workbook_15_kernel_density_estimation.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "workbook_15_kernel_density_estimation.ipynb",
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "20be466070cd43319544a4db1437d490": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_44e535fae72a4df498185164d7e0e181",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_e8151aea9ba841d2b3b44088a5325f29",
              "IPY_MODEL_57d30cafce30404dbb9a5c60524ba259"
            ]
          },
          "model_module_version": "1.5.0"
        },
        "44e535fae72a4df498185164d7e0e181": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "e8151aea9ba841d2b3b44088a5325f29": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DropdownModel",
          "state": {
            "_options_labels": [
              "gaussian",
              "tophat",
              "epanechnikov",
              "exponential",
              "linear",
              "cosine"
            ],
            "_view_name": "DropdownView",
            "style": "IPY_MODEL_f3240cd9a9d242e5b145e38d5a829531",
            "_dom_classes": [],
            "description": "Kernel:",
            "_model_name": "DropdownModel",
            "index": 0,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "disabled": false,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_c94389395b9d43a0b72469e18c0540dd"
          },
          "model_module_version": "1.5.0"
        },
        "57d30cafce30404dbb9a5c60524ba259": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "VBoxModel",
          "state": {
            "_view_name": "VBoxView",
            "_dom_classes": [],
            "_model_name": "VBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_de7cccfc43c84754a21ba9337490587f",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_dacbaf979527410e8079eeabe5d7c973",
              "IPY_MODEL_73c135b061544758a6b82850c41fafb7"
            ]
          },
          "model_module_version": "1.5.0"
        },
        "f3240cd9a9d242e5b145e38d5a829531": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          },
          "model_module_version": "1.5.0"
        },
        "c94389395b9d43a0b72469e18c0540dd": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "de7cccfc43c84754a21ba9337490587f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "dacbaf979527410e8079eeabe5d7c973": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatSliderModel",
          "state": {
            "_view_name": "FloatSliderView",
            "style": "IPY_MODEL_5822ed1760a447d4832a5a7b0012dea2",
            "_dom_classes": [],
            "description": "bandwidth",
            "step": 0.1,
            "_model_name": "FloatSliderModel",
            "orientation": "horizontal",
            "max": 5,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 2.0003000000000006,
            "_view_count": null,
            "disabled": false,
            "_view_module_version": "1.5.0",
            "min": 0.01,
            "continuous_update": true,
            "readout_format": ".2f",
            "description_tooltip": null,
            "readout": true,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_d73fc960016548f8a6c79b1df8e63692"
          },
          "model_module_version": "1.5.0"
        },
        "73c135b061544758a6b82850c41fafb7": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_04aed196c0da4888a3e9fd9b9d3c48f9",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_47b91572c0794be8886e47ff99d79344",
              "IPY_MODEL_558dfdd1c2b94f8884b69ca75b699109",
              "IPY_MODEL_1e5aa9699fd842339e157404ad468717"
            ]
          },
          "model_module_version": "1.5.0"
        },
        "5822ed1760a447d4832a5a7b0012dea2": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "SliderStyleModel",
          "state": {
            "_view_name": "StyleView",
            "handle_color": null,
            "_model_name": "SliderStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          },
          "model_module_version": "1.5.0"
        },
        "d73fc960016548f8a6c79b1df8e63692": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "04aed196c0da4888a3e9fd9b9d3c48f9": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "47b91572c0794be8886e47ff99d79344": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ButtonModel",
          "state": {
            "_view_name": "ButtonView",
            "style": "IPY_MODEL_ef67a15eb7a14d9f82c0680fcbc60744",
            "_dom_classes": [],
            "description": "Silverman",
            "_model_name": "ButtonModel",
            "button_style": "",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "tooltip": "",
            "_view_count": null,
            "disabled": false,
            "_view_module_version": "1.5.0",
            "layout": "IPY_MODEL_a77cb8d23c644ae18e5fb7a58f5d68be",
            "_model_module": "@jupyter-widgets/controls",
            "icon": ""
          },
          "model_module_version": "1.5.0"
        },
        "558dfdd1c2b94f8884b69ca75b699109": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ButtonModel",
          "state": {
            "_view_name": "ButtonView",
            "style": "IPY_MODEL_4d5e42c74ace4176932b88324cca16c9",
            "_dom_classes": [],
            "description": "ISJ",
            "_model_name": "ButtonModel",
            "button_style": "",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "tooltip": "",
            "_view_count": null,
            "disabled": false,
            "_view_module_version": "1.5.0",
            "layout": "IPY_MODEL_44b3cd51041c4e659ad6a5f98baf2c93",
            "_model_module": "@jupyter-widgets/controls",
            "icon": ""
          },
          "model_module_version": "1.5.0"
        },
        "1e5aa9699fd842339e157404ad468717": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ButtonModel",
          "state": {
            "_view_name": "ButtonView",
            "style": "IPY_MODEL_736dcd9a89344456bdc8064bc762f13c",
            "_dom_classes": [],
            "description": "Cross Validation",
            "_model_name": "ButtonModel",
            "button_style": "",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "tooltip": "",
            "_view_count": null,
            "disabled": false,
            "_view_module_version": "1.5.0",
            "layout": "IPY_MODEL_8735ad8208f84b36a3fa916c3caa7b2e",
            "_model_module": "@jupyter-widgets/controls",
            "icon": ""
          },
          "model_module_version": "1.5.0"
        },
        "ef67a15eb7a14d9f82c0680fcbc60744": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ButtonStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ButtonStyleModel",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "button_color": null,
            "font_weight": "",
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          },
          "model_module_version": "1.5.0"
        },
        "a77cb8d23c644ae18e5fb7a58f5d68be": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "4d5e42c74ace4176932b88324cca16c9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ButtonStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ButtonStyleModel",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "button_color": null,
            "font_weight": "",
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          },
          "model_module_version": "1.5.0"
        },
        "44b3cd51041c4e659ad6a5f98baf2c93": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "736dcd9a89344456bdc8064bc762f13c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ButtonStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ButtonStyleModel",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "button_color": null,
            "font_weight": "",
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          },
          "model_module_version": "1.5.0"
        },
        "8735ad8208f84b36a3fa916c3caa7b2e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        },
        "7de5fbb0b08b40f68b29d3b8d4f7af55": {
          "model_module": "@jupyter-widgets/output",
          "model_name": "OutputModel",
          "state": {
            "_view_name": "OutputView",
            "msg_id": "",
            "_dom_classes": [],
            "_model_name": "OutputModel",
            "outputs": [
              {
                "output_type": "display_data",
                "metadata": {
                  "tags": [],
                  "needs_background": "light"
                },
                "image/png": "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\n",
                "text/plain": "<Figure size 720x576 with 1 Axes>"
              }
            ],
            "_view_module": "@jupyter-widgets/output",
            "_model_module_version": "1.0.0",
            "_view_count": null,
            "_view_module_version": "1.0.0",
            "layout": "IPY_MODEL_2462f4a464474304bbf26643f7fab70e",
            "_model_module": "@jupyter-widgets/output"
          },
          "model_module_version": "1.0.0"
        },
        "2462f4a464474304bbf26643f7fab70e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          },
          "model_module_version": "1.2.0"
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/fbeilstein/machine_learning/blob/master/workbook_15_kernel_density_estimation.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#@title #Quiz\n",
        "\n",
        "%%html\n",
        "<link   rel=\"stylesheet\" href=\"https://fbeilstein.github.io/machine_learning/js_common/styles.css\"/>\n",
        "<script src=\"https://fbeilstein.github.io/machine_learning/js_common/script.js\"></script>\n",
        "<script>\n",
        "load_quiz_from(\"https://fbeilstein.github.io/machine_learning/lecture_15_kernel_density_estimation/quiz_1.txt\")\n",
        "</script>"
      ],
      "metadata": {
        "cellView": "form",
        "id": "bjgbPz3M8AJh",
        "outputId": "f6d4aa1c-1de3-4d0c-d23a-47a1a6be7661",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 247
        }
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<link   rel=\"stylesheet\" href=\"https://fbeilstein.github.io/machine_learning/js_common/styles.css\"/>\n",
              "<script src=\"https://fbeilstein.github.io/machine_learning/js_common/script.js\"></script>\n",
              "<script>\n",
              "load_quiz_from(\"https://fbeilstein.github.io/machine_learning/lecture_15_kernel_density_estimation/quiz_1.txt\")\n",
              "</script>\n"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-kqkuvkFAFll"
      },
      "source": [
        "#Bandwidth and kernel influence on 1D KDE"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DBOKZhb7oGvY"
      },
      "source": [
        "The following demo uses KDEpy library for Silverman's and ISJ bandwidth estimation methods.\n",
        "Start this code to install it:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8UvovqjzWjFO",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 238
        },
        "outputId": "9e986f6a-16a4-4806-cedf-9af95b435987"
      },
      "source": [
        "# install KDEpy library before starting the following demo\n",
        "!pip install KDEpy"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting KDEpy\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/d5/a2/8c07773c71ef2e6de288fd274cc874f876764ef834849e0418fa375e0123/KDEpy-1.0.5-cp36-cp36m-manylinux2010_x86_64.whl (396kB)\n",
            "\r\u001b[K     |▉                               | 10kB 15.8MB/s eta 0:00:01\r\u001b[K     |█▋                              | 20kB 2.1MB/s eta 0:00:01\r\u001b[K     |██▌                             | 30kB 2.9MB/s eta 0:00:01\r\u001b[K     |███▎                            | 40kB 2.9MB/s eta 0:00:01\r\u001b[K     |████▏                           | 51kB 2.5MB/s eta 0:00:01\r\u001b[K     |█████                           | 61kB 2.7MB/s eta 0:00:01\r\u001b[K     |█████▉                          | 71kB 3.0MB/s eta 0:00:01\r\u001b[K     |██████▋                         | 81kB 3.4MB/s eta 0:00:01\r\u001b[K     |███████▍                        | 92kB 3.5MB/s eta 0:00:01\r\u001b[K     |████████▎                       | 102kB 3.3MB/s eta 0:00:01\r\u001b[K     |█████████                       | 112kB 3.3MB/s eta 0:00:01\r\u001b[K     |██████████                      | 122kB 3.3MB/s eta 0:00:01\r\u001b[K     |██████████▊                     | 133kB 3.3MB/s eta 0:00:01\r\u001b[K     |███████████▋                    | 143kB 3.3MB/s eta 0:00:01\r\u001b[K     |████████████▍                   | 153kB 3.3MB/s eta 0:00:01\r\u001b[K     |█████████████▎                  | 163kB 3.3MB/s eta 0:00:01\r\u001b[K     |██████████████                  | 174kB 3.3MB/s eta 0:00:01\r\u001b[K     |██████████████▉                 | 184kB 3.3MB/s eta 0:00:01\r\u001b[K     |███████████████▊                | 194kB 3.3MB/s eta 0:00:01\r\u001b[K     |████████████████▌               | 204kB 3.3MB/s eta 0:00:01\r\u001b[K     |█████████████████▍              | 215kB 3.3MB/s eta 0:00:01\r\u001b[K     |██████████████████▏             | 225kB 3.3MB/s eta 0:00:01\r\u001b[K     |███████████████████             | 235kB 3.3MB/s eta 0:00:01\r\u001b[K     |███████████████████▉            | 245kB 3.3MB/s eta 0:00:01\r\u001b[K     |████████████████████▋           | 256kB 3.3MB/s eta 0:00:01\r\u001b[K     |█████████████████████▌          | 266kB 3.3MB/s eta 0:00:01\r\u001b[K     |██████████████████████▎         | 276kB 3.3MB/s eta 0:00:01\r\u001b[K     |███████████████████████▏        | 286kB 3.3MB/s eta 0:00:01\r\u001b[K     |████████████████████████        | 296kB 3.3MB/s eta 0:00:01\r\u001b[K     |████████████████████████▉       | 307kB 3.3MB/s eta 0:00:01\r\u001b[K     |█████████████████████████▋      | 317kB 3.3MB/s eta 0:00:01\r\u001b[K     |██████████████████████████▌     | 327kB 3.3MB/s eta 0:00:01\r\u001b[K     |███████████████████████████▎    | 337kB 3.3MB/s eta 0:00:01\r\u001b[K     |████████████████████████████    | 348kB 3.3MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████   | 358kB 3.3MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████▊  | 368kB 3.3MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████▋ | 378kB 3.3MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████▍| 389kB 3.3MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 399kB 3.3MB/s \n",
            "\u001b[?25hRequirement already satisfied: numpy>=1.14.2 in /usr/local/lib/python3.6/dist-packages (from KDEpy) (1.18.5)\n",
            "Requirement already satisfied: matplotlib>=2.2.0 in /usr/local/lib/python3.6/dist-packages (from KDEpy) (3.2.2)\n",
            "Requirement already satisfied: scipy>=1.0.1 in /usr/local/lib/python3.6/dist-packages (from KDEpy) (1.4.1)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=2.2.0->KDEpy) (1.2.0)\n",
            "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=2.2.0->KDEpy) (2.4.7)\n",
            "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=2.2.0->KDEpy) (2.8.1)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=2.2.0->KDEpy) (0.10.0)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.6/dist-packages (from python-dateutil>=2.1->matplotlib>=2.2.0->KDEpy) (1.15.0)\n",
            "Installing collected packages: KDEpy\n",
            "Successfully installed KDEpy-1.0.5\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2DTeu9JloX_n"
      },
      "source": [
        "Use the following demo to get the intuition on\n",
        "* How one-dimensional KDE works\n",
        "* How bandwidth influences the result\n",
        "* How good different methods (Silverman, ISJ, Cross validation) estimate the bandwidth"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Duqitq4pDd-J",
        "cellView": "form",
        "colab": {
          "resources": {
            "http://localhost:8080/nbextensions/google.colab/tabbar.css": {
              "data": "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",
              "ok": true,
              "headers": [
                [
                  "content-type",
                  "text/css"
                ]
              ],
              "status": 200,
              "status_text": ""
            },
            "http://localhost:8080/nbextensions/google.colab/tabbar_main.min.js": {
              "data": "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",
              "ok": true,
              "headers": [
                [
                  "content-type",
                  "application/javascript"
                ]
              ],
              "status": 200,
              "status_text": ""
            }
          },
          "base_uri": "https://localhost:8080/",
          "height": 674,
          "referenced_widgets": [
            "20be466070cd43319544a4db1437d490",
            "44e535fae72a4df498185164d7e0e181",
            "e8151aea9ba841d2b3b44088a5325f29",
            "57d30cafce30404dbb9a5c60524ba259",
            "f3240cd9a9d242e5b145e38d5a829531",
            "c94389395b9d43a0b72469e18c0540dd",
            "de7cccfc43c84754a21ba9337490587f",
            "dacbaf979527410e8079eeabe5d7c973",
            "73c135b061544758a6b82850c41fafb7",
            "5822ed1760a447d4832a5a7b0012dea2",
            "d73fc960016548f8a6c79b1df8e63692",
            "04aed196c0da4888a3e9fd9b9d3c48f9",
            "47b91572c0794be8886e47ff99d79344",
            "558dfdd1c2b94f8884b69ca75b699109",
            "1e5aa9699fd842339e157404ad468717",
            "ef67a15eb7a14d9f82c0680fcbc60744",
            "a77cb8d23c644ae18e5fb7a58f5d68be",
            "4d5e42c74ace4176932b88324cca16c9",
            "44b3cd51041c4e659ad6a5f98baf2c93",
            "736dcd9a89344456bdc8064bc762f13c",
            "8735ad8208f84b36a3fa916c3caa7b2e",
            "7de5fbb0b08b40f68b29d3b8d4f7af55",
            "2462f4a464474304bbf26643f7fab70e"
          ]
        },
        "outputId": "558707ea-40a5-456c-fca0-ab6c3f8f051b"
      },
      "source": [
        "#@title Bandwidth and kernel in KDE\n",
        "from google.colab import widgets\n",
        "import ipywidgets\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import scipy\n",
        "from sklearn.neighbors import KernelDensity\n",
        "from sklearn.model_selection import learning_curve, GridSearchCV\n",
        "from sklearn.model_selection import LeaveOneOut #train_test_split\n",
        "from KDEpy import FFTKDE\n",
        "import IPython\n",
        "from google.colab import output\n",
        "\n",
        "\n",
        "main_str = '''\n",
        "<html>\n",
        "\n",
        "<head>\n",
        "  <script src=\"https://cdn.plot.ly/plotly-latest.min.js\"></script>\n",
        "\n",
        "  <script type=\"application/javascript\">\n",
        "    var xs = [];\n",
        "    var ys = [];\n",
        "    var x_min = -5.0;\n",
        "    var y_min =  0.0;\n",
        "    var x_max = +5.0;\n",
        "    var y_max = +0.8;\n",
        "    for (var x = x_min; x <= x_max; x += 0.01)\n",
        "      xs.push(x);\n",
        "    for (var y = y_min; y <= y_max; y += 0.1)\n",
        "      ys.push(y);\n",
        "    var colors = ['blue', 'green', 'red'];\n",
        "    var coefs = [1.0/3.0, 1.0/3.0, 1.0/3.0];\n",
        "    var rand_pts = [];\n",
        "\n",
        "   function random_point(mu, sigma) {\n",
        "     var z = Math.sqrt(-2.0 * Math.log(Math.random())) * Math.cos(2.0 * Math.PI * Math.random());\n",
        "     return z * sigma + mu;\n",
        "   }\n",
        "\n",
        "    function loss_function_2d() {\n",
        "      var canvas = document.getElementById(\"paint\");\n",
        "      var w = canvas.width;\n",
        "      var h = canvas.height;\n",
        "\n",
        "      //var size = adjust_canvas();\n",
        "      //var w = size[0];\n",
        "      //var h = size[1];\n",
        "\n",
        "      mu = [];\n",
        "      sigma = [];\n",
        "      C = Math.sqrt(2 * Math.PI);\n",
        "      for (var i = 0; i < pts.length; ++i) {\n",
        "          var x = pts[i][0] / w * (x_max - x_min) + x_min;\n",
        "        var y = (1.0 - pts[i][1] / h) * (y_max - y_min) + y_min;\n",
        "          mu.push(x);\n",
        "          sigma.push(1.0 / (y * C));\n",
        "      }\n",
        "\n",
        "      data = [];\n",
        "      for (var trace_idx = 0; trace_idx < pts.length; ++trace_idx) {\n",
        "          var trace = [];\n",
        "          for (var i = 0; i < xs.length; ++i) {\n",
        "              var delta = (xs[i] - mu[trace_idx]) / sigma[trace_idx];\n",
        "              trace.push(Math.exp(-delta*delta / 2.0) / sigma[trace_idx] / C);\n",
        "          }\n",
        "          data.push({\n",
        "            x: xs,\n",
        "            y: trace,\n",
        "            showscale: false,\n",
        "            opacity: coefs[trace_idx],\n",
        "            type: 'scatter',\n",
        "            mode: 'line',\n",
        "            name: 'gaussian ' + trace_idx,\n",
        "            line: {\n",
        "              width: 2,\n",
        "              color: colors[trace_idx]\n",
        "            },\n",
        "          });\n",
        "      }\n",
        "\n",
        "      var sum_trace = [];\n",
        "      for (var i = 0; i < xs.length; ++i) {\n",
        "          var sum = 0.0;\n",
        "          for (var trace_idx = 0; trace_idx < pts.length; ++trace_idx) {\n",
        "              var delta = (xs[i] - mu[trace_idx]) / sigma[trace_idx];\n",
        "              sum += coefs[trace_idx] * Math.exp(-delta*delta / 2.0) / sigma[trace_idx] / C;\n",
        "          }\n",
        "          sum_trace.push(sum);\n",
        "      }\n",
        "      data.push({\n",
        "            x: xs,\n",
        "            y: sum_trace,\n",
        "            showscale: false,\n",
        "            opacity: coefs[trace_idx],\n",
        "            type: 'scatter',\n",
        "            mode: 'line',\n",
        "            name: 'sum',\n",
        "            line: {\n",
        "              width: 2,\n",
        "              color: 'black'\n",
        "            },\n",
        "          });\n",
        "\n",
        "      var N_points = 200;\n",
        "      rand_pts = [];\n",
        "      var ys = [];\n",
        "      for (var i = 0; i < N_points; ++i) {\n",
        "        active_gaussian = 0;\n",
        "        var rnd = Math.random();\n",
        "        if (rnd > coefs[0])\n",
        "            active_gaussian = 1;\n",
        "        if (rnd > coefs[0] + coefs[1])\n",
        "            active_gaussian = 2;\n",
        "        //active_gaussian = Math.floor(Math.random() * Math.floor(3));\n",
        "        rand_pts.push(random_point(mu[active_gaussian], sigma[active_gaussian]));\n",
        "        ys.push(0.0);\n",
        "      }\n",
        "      data.push({\n",
        "            x: rand_pts,\n",
        "            y: ys,\n",
        "            showscale: false,\n",
        "            opacity: 1.0,\n",
        "            type: 'scatter',\n",
        "            mode: 'markers',\n",
        "            name: 'random datapoints',\n",
        "            marker: {\n",
        "              symbol: 'line-ns',\n",
        "              size: 20,\n",
        "              line :{\n",
        "                width: 1,\n",
        "                color: 'black'\n",
        "              }\n",
        "            }\n",
        "          });\n",
        "\n",
        "      var cell_for_plot = document.getElementById(\"for_right_plot\");\n",
        "      var layout = {\n",
        "          //title: 'Error Function',\n",
        "          xaxis: {'range': [x_min, x_max], 'domain': [0.0, 1.0], 'fixedrange': true},\n",
        "          yaxis: {'range': [y_min, y_max], 'domain': [0.0, 1.0], 'fixedrange': true},\n",
        "          width: 800,\n",
        "          height: 500,\n",
        "          autosize: false,\n",
        "          showlegend: true,\n",
        "          margin: {\n",
        "            l: 50,\n",
        "            r: 50,\n",
        "            b: 50,\n",
        "            t: 50\n",
        "          }\n",
        "        };\n",
        "      Plotly.newPlot(cell_for_plot, data, layout);\n",
        "    }\n",
        "\n",
        "    var pts = [[100, 300], [300, 300], [500, 300]];\n",
        "    var active_idx = -1;\n",
        "\n",
        "    function mdown_handle(event) {\n",
        "      //console.log('down');\n",
        "      event.stopPropagation();\n",
        "      x = event.offsetX;\n",
        "      y = event.offsetY;\n",
        "      for (var i = 0; i < pts.length; ++i)\n",
        "        if ((x - pts[i][0])*(x - pts[i][0]) + (y - pts[i][1])*(y - pts[i][1]) < 500)\n",
        "            active_idx = i;\n",
        "      //console.log('ACTIVE', active_idx);\n",
        "    }\n",
        "\n",
        "    function mmove_handle(event) {\n",
        "      event.stopPropagation();\n",
        "      //console.log('move');\n",
        "      if (active_idx < 0)\n",
        "          return;\n",
        "      pts[active_idx] = [event.offsetX, event.offsetY];\n",
        "      loss_function_2d();\n",
        "    }\n",
        "\n",
        "    function mup_handle(event) {\n",
        "      active_idx = -1;\n",
        "    }\n",
        "\n",
        "    async function call_once() {\n",
        "      const result = await google.colab.kernel.invokeFunction('notebook.redraw', [coefs, mu, sigma, rand_pts], {});\n",
        "      //const text = result.data['application/json'];\n",
        "    }\n",
        "    var old_state = false;\n",
        "    function adjust_canvas() {\n",
        "      var plt_div = document.getElementById('for_right_plot');\n",
        "      var el = plt_div.getElementsByClassName('xy')[0];\n",
        "      if (!el)\n",
        "          return [600, 600];\n",
        "      var rect = el.getBoundingClientRect();\n",
        "      var canvas = document.getElementById(\"paint\");\n",
        "      //console.log(rect);\n",
        "      canvas.style.cssText = \"position:absolute; top:\" + parseInt(rect.top)\n",
        "                          + \"px; left: \" + parseInt(rect.left)\n",
        "                          + \"px; width:\" + parseInt(rect.width)\n",
        "                          + \"px; height:\" + parseInt(rect.height)\n",
        "                          + \"px; z-index:1000;\";\n",
        "      canvas.width = rect.width;\n",
        "      canvas.height = rect.height;\n",
        "      var ctx = canvas.getContext('2d');\n",
        "      ctx.clearRect(0, 0, canvas.width, canvas.height);\n",
        "\n",
        "      if ((canvas.offsetParent === null) != old_state)\n",
        "        call_once();\n",
        "      old_state = (canvas.offsetParent === null);\n",
        "\n",
        "      //loss_function_2d();\n",
        "      return [canvas.width, canvas.height];\n",
        "    }\n",
        "\n",
        "    function draw_slider() {\n",
        "        var canvas = document.getElementById(\"slider\");\n",
        "        var ctx = canvas.getContext('2d');\n",
        "        var h = canvas.height;\n",
        "        var w = canvas.width;\n",
        "        var cur_start = 0;\n",
        "        ctx.clearRect(0, 0, canvas.width, canvas.height);\n",
        "        for (var i = 0; i < pts.length; ++i) {\n",
        "            ctx.fillStyle = colors[i];\n",
        "            ctx.fillRect(0, cur_start, w, h*coefs[i]);\n",
        "            cur_start += h*coefs[i];\n",
        "        }\n",
        "\n",
        "    }\n",
        "\n",
        "    var active_coef = -1;\n",
        "    function mdown(event) {\n",
        "    //console.log('ddd');\n",
        "      event.stopPropagation();\n",
        "      x = event.offsetX;\n",
        "      y = event.offsetY;\n",
        "      var canvas = document.getElementById(\"slider\");\n",
        "      var h = canvas.height;\n",
        "      var H = 0;\n",
        "      for (var i = 0; i < coefs.length - 1; ++i) {\n",
        "        H += coefs[i] * h;\n",
        "        if (y > H - 10 && y < H +1)\n",
        "            active_coef = i;\n",
        "      }\n",
        "      //console.log(active_coef, coefs);\n",
        "    }\n",
        "\n",
        "    function mmove(event) { // for double-slider only\n",
        "      event.stopPropagation();\n",
        "      if (active_coef < 0)\n",
        "        return;\n",
        "      x = event.offsetX;\n",
        "      y = event.offsetY;\n",
        "      var canvas = document.getElementById(\"slider\");\n",
        "      var h = canvas.height;\n",
        "      var new_sum = y / h;\n",
        "      if (active_coef == 0) {\n",
        "        if (coefs[0] + coefs[1] < new_sum)\n",
        "            coefs[1] = 0.0;\n",
        "        else\n",
        "            coefs[1] = coefs[0] + coefs[1] - new_sum;\n",
        "        coefs[0] = new_sum;\n",
        "        coefs[2] = 1.0 - coefs[0] - coefs[1];\n",
        "      }\n",
        "      if (active_coef == 1) {\n",
        "        if (coefs[0] > new_sum)\n",
        "            coefs[0] = new_sum;\n",
        "        coefs[1] = new_sum - coefs[0];\n",
        "        coefs[2] = 1.0 - coefs[0] - coefs[1];\n",
        "      }\n",
        "      draw_slider();\n",
        "      loss_function_2d();\n",
        "    }\n",
        "\n",
        "    function mup(event) {\n",
        "      active_coef = -1;\n",
        "    }\n",
        "\n",
        "\n",
        "  </script>\n",
        "</head>\n",
        "\n",
        "<body>\n",
        "\n",
        "<canvas id=\"paint\" width=600 height=600\n",
        "    onmousedown=\"mdown_handle(event)\"\n",
        "    onmousemove=\"mmove_handle(event)\"\n",
        "    onmouseup=\"mup_handle(event)\">\n",
        "</canvas>\n",
        "\n",
        "<canvas id=\"slider\" width=40 height=250\n",
        "    onmousedown=\"mdown(event)\"\n",
        "    onmousemove=\"mmove(event)\"\n",
        "    onmouseup=\"mup(event)\"\n",
        "    style=\"position:absolute;left:700px;top:250px;width:40px;height:250px;z-index:1000;\">\n",
        "</canvas>\n",
        "\n",
        "\n",
        "<div id=\"everything\" style=\"height:600px; width:600px;\">\n",
        "  <div id=\"for_right_plot\">\n",
        "  </div>\n",
        "</div>\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "</body>\n",
        "\n",
        "<script>\n",
        "function init()\n",
        "{\n",
        "  loss_function_2d();\n",
        "  adjust_canvas();\n",
        "  loss_function_2d();\n",
        "  setInterval(adjust_canvas, 1000);\n",
        "  draw_slider();\n",
        "}\n",
        "init();\n",
        "</script>\n",
        "\n",
        "</html>\n",
        "'''\n",
        "\n",
        "params = []\n",
        "X = []\n",
        "b_w = ipywidgets.FloatSlider(description='bandwidth', min=0.01, max=5.0, value=2.0)\n",
        "ker_type = ipywidgets.Dropdown(\n",
        "  options=['gaussian', 'tophat', 'epanechnikov', 'exponential','linear','cosine'],\n",
        "  value='gaussian',\n",
        "  description='Kernel:',\n",
        "  disabled=False)\n",
        "\n",
        "def set_slider(x):\n",
        "  bandwidths = np.linspace(b_w.min, b_w.max, 100)\n",
        "  loo = LeaveOneOut()\n",
        "  grid = GridSearchCV(KernelDensity(kernel=ker_type.value), {'bandwidth': bandwidths}, cv=loo)\n",
        "  grid.fit(np.array(X)[:, None]);\n",
        "  b_w.value = grid.best_params_['bandwidth']\n",
        "\n",
        "def set_slider2(x):\n",
        "  ker = {'gaussian':'gaussian', 'cosine':'cosine',\n",
        "         'tophat':'box', 'epanechnikov':'epa',\n",
        "         'exponential':'exponential', 'linear':'tri'}\n",
        "  kde = FFTKDE(bw='ISJ', kernel=ker[ker_type.value])\n",
        "  kde.fit(X)(None)\n",
        "  b_w.value = kde.bw\n",
        "\n",
        "def set_slider1(x):\n",
        "  #std = np.sqrt(np.var(X, axis=0))\n",
        "  #silverman = ((4/(3* len(np.array(X)[:, None])))*std**5)**(1/5)\n",
        "  #b_w.value = silverman\n",
        "\n",
        "  ker = {'gaussian':'gaussian', 'cosine':'cosine',\n",
        "         'tophat':'box', 'epanechnikov':'epa',\n",
        "         'exponential':'exponential', 'linear':'tri'}\n",
        "  kde = FFTKDE(bw='silverman', kernel=ker[ker_type.value])\n",
        "  kde.fit(X)(None)\n",
        "  b_w.value = kde.bw\n",
        "\n",
        "def plot_pdf(plot_separate, k_d = None):\n",
        "  #return\n",
        "  global params\n",
        "  global X\n",
        "  if not params or not X:\n",
        "    return\n",
        "\n",
        "  pi1, pi2, pi3, mu1, sig1, mu2, sig2, mu3, sig3 = params\n",
        "  x = np.linspace(np.min(X), np.max(X), 1001)\n",
        "  fig = plt.figure(figsize=(10, 8))\n",
        "  P_sum = np.zeros(x.shape)\n",
        "  for mu, sig, c, pi, i in [[mu1, sig1, 'r', pi1, 1], [mu2, sig2, 'g', pi2, 2], [mu3, sig3, 'b', pi3, 3]]:\n",
        "    P = scipy.stats.norm.pdf(x, mu, sig)\n",
        "    P_sum += P * pi\n",
        "    if plot_separate:\n",
        "      plt.plot(x, P, c, linewidth=2.0, label=\"{}\".format(i), alpha=np.clip(pi, 0.0, 1.0))\n",
        "\n",
        "  plt.plot(x, P_sum, 'k', linewidth=3.0, label=\"mixture\")\n",
        "  plt.plot(np.array(X), np.full_like(np.array(X), 0.0), '|k', markeredgewidth=1)\n",
        "  plt.title('True pdf / pdf estimate')\n",
        "  plt.xlabel('x', fontsize=18)\n",
        "  plt.ylabel('pdf', fontsize=18)\n",
        "  plt.legend()\n",
        "  #plt.show()\n",
        "\n",
        "\n",
        "def ker_plot(b_w, k_t):\n",
        "  global params\n",
        "  global X\n",
        "  if not params or not X:\n",
        "    return\n",
        "  kde = KernelDensity(bandwidth=b_w, kernel=k_t)\n",
        "  kde.fit(np.array(X)[:, None])\n",
        "  x_d = np.linspace(np.min(X), np.max(X), 1001)\n",
        "  logprob = kde.score_samples(x_d[:, None])\n",
        "  plot_pdf(False)\n",
        "  plt.fill_between(x_d, np.exp(logprob), alpha=0.5, color='r')\n",
        "\n",
        "def redraw(coefs, mu, sigma, points):\n",
        "  global X\n",
        "  global params\n",
        "  X = points\n",
        "  params = [coefs[0], coefs[1], coefs[2],\n",
        "            mu[0], sigma[0],\n",
        "            mu[1], sigma[1],\n",
        "            mu[2], sigma[2]]\n",
        "  #plot_pdf(False)\n",
        "  b_w.value += 0.0001\n",
        "\n",
        "\n",
        "tb = widgets.TabBar(['mix distribution', 'find KDE'], location='top')\n",
        "with tb.output_to(0):\n",
        "  display(IPython.display.HTML(main_str))\n",
        "  output.register_callback('notebook.redraw', redraw)\n",
        "\n",
        "with tb.output_to(1):\n",
        "  btn = ipywidgets.Button(\n",
        "    description='Cross Validation',\n",
        "    disabled=False,\n",
        "    button_style='')\n",
        "  btn1 = ipywidgets.Button(\n",
        "    description='Silverman',\n",
        "    disabled=False,\n",
        "    button_style='')\n",
        "  btn2 = ipywidgets.Button(\n",
        "    description='ISJ',\n",
        "    disabled=False,\n",
        "    button_style='')\n",
        "  out = ipywidgets.interactive_output(ker_plot, {\"b_w\": b_w, \"k_t\": ker_type})\n",
        "  display(ipywidgets.HBox([ker_type, ipywidgets.VBox([b_w, ipywidgets.HBox([btn1, btn2, btn])])]), out)\n",
        "  btn.on_click(set_slider)\n",
        "  btn1.on_click(set_slider1)\n",
        "  btn2.on_click(set_slider2)\n",
        "  b_w.value += 0.0001\n",
        "\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<link rel=stylesheet type=text/css href='/nbextensions/google.colab/tabbar.css'></link>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": [
              "outputarea_id3"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<script src='/nbextensions/google.colab/tabbar_main.min.js'></script>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": [
              "outputarea_id3"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<div id=\"id3\"></div>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": [
              "outputarea_id3"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "window[\"0f0060fe-f7ad-11ea-9b9e-0242ac1c0002\"] = colab_lib.createTabBar({\"location\": \"top\", \"elementId\": \"id3\", \"tabNames\": [\"mix distribution\", \"find KDE\"], \"initialSelection\": 0, \"contentBorder\": [\"0px\"], \"contentHeight\": [\"initial\"], \"borderColor\": [\"#a7a7a7\"]});\n",
              "//# sourceURL=js_8ba1890bcf"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": [
              "outputarea_id3"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "window[\"0f00b07c-f7ad-11ea-9b9e-0242ac1c0002\"] = window[\"id3\"].setSelectedTabIndex(0);\n",
              "//# sourceURL=js_4086314917"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": [
              "outputarea_id3"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "window[\"0f01ab44-f7ad-11ea-9b9e-0242ac1c0002\"] = google.colab.output.getActiveOutputArea();\n",
              "//# sourceURL=js_82725e5288"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": [
              "id3_content_0",
              "outputarea_id3"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "window[\"0f01e5f0-f7ad-11ea-9b9e-0242ac1c0002\"] = document.querySelector(\"#id3_content_0\");\n",
              "//# sourceURL=js_4023db1f0c"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": [
              "id3_content_0",
              "outputarea_id3"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "window[\"0f021eda-f7ad-11ea-9b9e-0242ac1c0002\"] = google.colab.output.setActiveOutputArea(window[\"0f01e5f0-f7ad-11ea-9b9e-0242ac1c0002\"]);\n",
              "//# sourceURL=js_328a992244"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": [
              "id3_content_0",
              "outputarea_id3"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "window[\"0f025f30-f7ad-11ea-9b9e-0242ac1c0002\"] = window[\"id3\"].setSelectedTabIndex(0);\n",
              "//# sourceURL=js_1b097e43bb"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": [
              "id3_content_0",
              "outputarea_id3"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "<html>\n",
              "\n",
              "<head>\n",
              "  <script src=\"https://cdn.plot.ly/plotly-latest.min.js\"></script> \n",
              "\n",
              "  <script type=\"application/javascript\">\n",
              "    var xs = [];\n",
              "    var ys = [];\n",
              "    var x_min = -5.0;\n",
              "    var y_min =  0.0;\n",
              "    var x_max = +5.0;\n",
              "    var y_max = +0.8;\n",
              "    for (var x = x_min; x <= x_max; x += 0.01)\n",
              "      xs.push(x);\n",
              "    for (var y = y_min; y <= y_max; y += 0.1)\n",
              "      ys.push(y);\n",
              "    var colors = ['blue', 'green', 'red'];\n",
              "    var coefs = [1.0/3.0, 1.0/3.0, 1.0/3.0];\n",
              "    var rand_pts = [];\n",
              "\n",
              "   function random_point(mu, sigma) {\n",
              "     var z = Math.sqrt(-2.0 * Math.log(Math.random())) * Math.cos(2.0 * Math.PI * Math.random());\n",
              "     return z * sigma + mu;\n",
              "   }\n",
              "\n",
              "    function loss_function_2d() {\n",
              "      var canvas = document.getElementById(\"paint\");\n",
              "      var w = canvas.width;\n",
              "      var h = canvas.height;\n",
              "\n",
              "      //var size = adjust_canvas();\n",
              "      //var w = size[0];\n",
              "      //var h = size[1];\n",
              "\n",
              "      mu = [];\n",
              "      sigma = [];\n",
              "      C = Math.sqrt(2 * Math.PI);\n",
              "      for (var i = 0; i < pts.length; ++i) {\n",
              "          var x = pts[i][0] / w * (x_max - x_min) + x_min;\n",
              "        var y = (1.0 - pts[i][1] / h) * (y_max - y_min) + y_min;\n",
              "          mu.push(x);\n",
              "          sigma.push(1.0 / (y * C));\n",
              "      }\n",
              "\n",
              "      data = [];\n",
              "      for (var trace_idx = 0; trace_idx < pts.length; ++trace_idx) {\n",
              "          var trace = [];\n",
              "          for (var i = 0; i < xs.length; ++i) {\n",
              "              var delta = (xs[i] - mu[trace_idx]) / sigma[trace_idx]; \n",
              "              trace.push(Math.exp(-delta*delta / 2.0) / sigma[trace_idx] / C);\n",
              "          }\n",
              "          data.push({\n",
              "            x: xs,\n",
              "            y: trace,\n",
              "            showscale: false, \n",
              "            opacity: coefs[trace_idx], \n",
              "            type: 'scatter',\n",
              "            mode: 'line',\n",
              "            name: 'gaussian ' + trace_idx,\n",
              "            line: {\n",
              "              width: 2,\n",
              "              color: colors[trace_idx]\n",
              "            },\n",
              "          });\n",
              "      }\n",
              "\n",
              "      var sum_trace = [];\n",
              "      for (var i = 0; i < xs.length; ++i) {\n",
              "          var sum = 0.0;\n",
              "          for (var trace_idx = 0; trace_idx < pts.length; ++trace_idx) {\n",
              "              var delta = (xs[i] - mu[trace_idx]) / sigma[trace_idx]; \n",
              "              sum += coefs[trace_idx] * Math.exp(-delta*delta / 2.0) / sigma[trace_idx] / C;\n",
              "          }\n",
              "          sum_trace.push(sum);\n",
              "      }\n",
              "      data.push({\n",
              "            x: xs,\n",
              "            y: sum_trace,\n",
              "            showscale: false, \n",
              "            opacity: coefs[trace_idx], \n",
              "            type: 'scatter',\n",
              "            mode: 'line',\n",
              "            name: 'sum',\n",
              "            line: {\n",
              "              width: 2,\n",
              "              color: 'black'\n",
              "            },\n",
              "          });\n",
              "\n",
              "      var N_points = 200;\n",
              "      rand_pts = [];\n",
              "      var ys = [];\n",
              "      for (var i = 0; i < N_points; ++i) {\n",
              "        active_gaussian = 0;\n",
              "        var rnd = Math.random();\n",
              "        if (rnd > coefs[0])\n",
              "            active_gaussian = 1;\n",
              "        if (rnd > coefs[0] + coefs[1])\n",
              "            active_gaussian = 2;\n",
              "        //active_gaussian = Math.floor(Math.random() * Math.floor(3));\n",
              "        rand_pts.push(random_point(mu[active_gaussian], sigma[active_gaussian]));\n",
              "        ys.push(0.0);\n",
              "      }\n",
              "      data.push({\n",
              "            x: rand_pts,\n",
              "            y: ys,\n",
              "            showscale: false, \n",
              "            opacity: 1.0, \n",
              "            type: 'scatter',\n",
              "            mode: 'markers',\n",
              "            name: 'random datapoints',\n",
              "            marker: {\n",
              "              symbol: 'line-ns',\n",
              "              size: 20,\n",
              "              line :{\n",
              "                width: 1,\n",
              "                color: 'black'\n",
              "              }\n",
              "            }\n",
              "          });\n",
              "\n",
              "      var cell_for_plot = document.getElementById(\"for_right_plot\");\n",
              "      var layout = {\n",
              "          //title: 'Error Function',\n",
              "          xaxis: {'range': [x_min, x_max], 'domain': [0.0, 1.0], 'fixedrange': true},\n",
              "          yaxis: {'range': [y_min, y_max], 'domain': [0.0, 1.0], 'fixedrange': true},\n",
              "          width: 800,\n",
              "          height: 500,\n",
              "          autosize: false,\n",
              "          showlegend: true,\n",
              "          margin: {\n",
              "            l: 50,\n",
              "            r: 50,\n",
              "            b: 50,\n",
              "            t: 50\n",
              "          }\n",
              "        };\n",
              "      Plotly.newPlot(cell_for_plot, data, layout);\n",
              "    }\n",
              "\n",
              "    var pts = [[100, 300], [300, 300], [500, 300]];\n",
              "    var active_idx = -1;\n",
              "\n",
              "    function mdown_handle(event) {\n",
              "      //console.log('down');\n",
              "      event.stopPropagation();\n",
              "      x = event.offsetX;\n",
              "      y = event.offsetY;\n",
              "      for (var i = 0; i < pts.length; ++i)\n",
              "        if ((x - pts[i][0])*(x - pts[i][0]) + (y - pts[i][1])*(y - pts[i][1]) < 500)\n",
              "            active_idx = i;\n",
              "      //console.log('ACTIVE', active_idx);\n",
              "    }\n",
              "\n",
              "    function mmove_handle(event) {\n",
              "      event.stopPropagation();\n",
              "      //console.log('move');\n",
              "      if (active_idx < 0)\n",
              "          return;\n",
              "      pts[active_idx] = [event.offsetX, event.offsetY];\n",
              "      loss_function_2d();\n",
              "    }\n",
              "\n",
              "    function mup_handle(event) {\n",
              "      active_idx = -1;\n",
              "    }\n",
              "\n",
              "    async function call_once() {\n",
              "      const result = await google.colab.kernel.invokeFunction('notebook.redraw', [coefs, mu, sigma, rand_pts], {});\n",
              "      //const text = result.data['application/json'];\n",
              "    }\n",
              "    var old_state = false;\n",
              "    function adjust_canvas() {\n",
              "      var plt_div = document.getElementById('for_right_plot');\n",
              "      var el = plt_div.getElementsByClassName('xy')[0];\n",
              "      if (!el)\n",
              "          return [600, 600];\n",
              "      var rect = el.getBoundingClientRect();\n",
              "      var canvas = document.getElementById(\"paint\");\n",
              "      //console.log(rect);\n",
              "      canvas.style.cssText = \"position:absolute; top:\" + parseInt(rect.top)\n",
              "                          + \"px; left: \" + parseInt(rect.left)\n",
              "                          + \"px; width:\" + parseInt(rect.width)\n",
              "                          + \"px; height:\" + parseInt(rect.height)\n",
              "                          + \"px; z-index:1000;\";\n",
              "      canvas.width = rect.width;\n",
              "      canvas.height = rect.height;\n",
              "      var ctx = canvas.getContext('2d');\n",
              "      ctx.clearRect(0, 0, canvas.width, canvas.height);\n",
              "       \n",
              "      if ((canvas.offsetParent === null) != old_state)\n",
              "        call_once();\n",
              "      old_state = (canvas.offsetParent === null);\n",
              "\n",
              "      //loss_function_2d();\n",
              "      return [canvas.width, canvas.height];\n",
              "    }\n",
              "\n",
              "    function draw_slider() {\n",
              "        var canvas = document.getElementById(\"slider\");\n",
              "        var ctx = canvas.getContext('2d');\n",
              "        var h = canvas.height;\n",
              "        var w = canvas.width;\n",
              "        var cur_start = 0;\n",
              "        ctx.clearRect(0, 0, canvas.width, canvas.height);\n",
              "        for (var i = 0; i < pts.length; ++i) {\n",
              "            ctx.fillStyle = colors[i];\n",
              "            ctx.fillRect(0, cur_start, w, h*coefs[i]);\n",
              "            cur_start += h*coefs[i];\n",
              "        }\n",
              "        \n",
              "    }\n",
              "\n",
              "    var active_coef = -1;\n",
              "    function mdown(event) {\n",
              "    //console.log('ddd');\n",
              "      event.stopPropagation();\n",
              "      x = event.offsetX;\n",
              "      y = event.offsetY;\n",
              "      var canvas = document.getElementById(\"slider\");\n",
              "      var h = canvas.height;\n",
              "      var H = 0;\n",
              "      for (var i = 0; i < coefs.length - 1; ++i) {\n",
              "        H += coefs[i] * h;\n",
              "        if (y > H - 10 && y < H +1)\n",
              "            active_coef = i;\n",
              "      }\n",
              "      //console.log(active_coef, coefs);\n",
              "    }\n",
              "\n",
              "    function mmove(event) { // for double-slider only\n",
              "      event.stopPropagation();\n",
              "      if (active_coef < 0)\n",
              "        return;\n",
              "      x = event.offsetX;\n",
              "      y = event.offsetY;\n",
              "      var canvas = document.getElementById(\"slider\");\n",
              "      var h = canvas.height;\n",
              "      var new_sum = y / h;\n",
              "      if (active_coef == 0) {\n",
              "        if (coefs[0] + coefs[1] < new_sum)\n",
              "            coefs[1] = 0.0;\n",
              "        else\n",
              "            coefs[1] = coefs[0] + coefs[1] - new_sum;\n",
              "        coefs[0] = new_sum;\n",
              "        coefs[2] = 1.0 - coefs[0] - coefs[1];\n",
              "      }\n",
              "      if (active_coef == 1) {\n",
              "        if (coefs[0] > new_sum)\n",
              "            coefs[0] = new_sum;\n",
              "        coefs[1] = new_sum - coefs[0];\n",
              "        coefs[2] = 1.0 - coefs[0] - coefs[1];\n",
              "      }\n",
              "      draw_slider();\n",
              "      loss_function_2d();\n",
              "    }\n",
              "\n",
              "    function mup(event) {\n",
              "      active_coef = -1;\n",
              "    }\n",
              "\n",
              "\n",
              "  </script>\n",
              "</head>\n",
              "\n",
              "<body>\n",
              "\n",
              "<canvas id=\"paint\" width=600 height=600\n",
              "    onmousedown=\"mdown_handle(event)\"\n",
              "    onmousemove=\"mmove_handle(event)\"\n",
              "    onmouseup=\"mup_handle(event)\">\n",
              "</canvas>\n",
              "\n",
              "<canvas id=\"slider\" width=40 height=250\n",
              "    onmousedown=\"mdown(event)\"\n",
              "    onmousemove=\"mmove(event)\"\n",
              "    onmouseup=\"mup(event)\"\n",
              "    style=\"position:absolute;left:700px;top:250px;width:40px;height:250px;z-index:1000;\">\n",
              "</canvas>\n",
              "\n",
              "\n",
              "<div id=\"everything\" style=\"height:600px; width:600px;\">\n",
              "  <div id=\"for_right_plot\">\n",
              "  </div>\n",
              "</div>\n",
              "\n",
              "\n",
              "\n",
              "\n",
              "</body>\n",
              "\n",
              "<script>\n",
              "function init()\n",
              "{\n",
              "  loss_function_2d();\n",
              "  adjust_canvas();\n",
              "  loss_function_2d();\n",
              "  setInterval(adjust_canvas, 1000);\n",
              "  draw_slider();\n",
              "}\n",
              "init();\n",
              "</script>\n",
              "\n",
              "</html>\n"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": [
              "id3_content_0",
              "outputarea_id3",
              "user_output"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "window[\"0f033ffe-f7ad-11ea-9b9e-0242ac1c0002\"] = google.colab.output.setActiveOutputArea(window[\"0f01ab44-f7ad-11ea-9b9e-0242ac1c0002\"]);\n",
              "//# sourceURL=js_a966006f2e"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": [
              "id3_content_0",
              "outputarea_id3"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "window[\"0f047dd8-f7ad-11ea-9b9e-0242ac1c0002\"] = google.colab.output.getActiveOutputArea();\n",
              "//# sourceURL=js_13b5210992"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": [
              "id3_content_1",
              "outputarea_id3"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "window[\"0f04bc62-f7ad-11ea-9b9e-0242ac1c0002\"] = document.querySelector(\"#id3_content_1\");\n",
              "//# sourceURL=js_0f873afd1f"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": [
              "id3_content_1",
              "outputarea_id3"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "window[\"0f051202-f7ad-11ea-9b9e-0242ac1c0002\"] = google.colab.output.setActiveOutputArea(window[\"0f04bc62-f7ad-11ea-9b9e-0242ac1c0002\"]);\n",
              "//# sourceURL=js_b8c4e123bb"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": [
              "id3_content_1",
              "outputarea_id3"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "window[\"0f054eca-f7ad-11ea-9b9e-0242ac1c0002\"] = window[\"id3\"].setSelectedTabIndex(1);\n",
              "//# sourceURL=js_e85b46f667"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": [
              "id3_content_1",
              "outputarea_id3"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "20be466070cd43319544a4db1437d490",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(Dropdown(description='Kernel:', options=('gaussian', 'tophat', 'epanechnikov', 'exponential', '…"
            ]
          },
          "metadata": {
            "tags": [
              "id3_content_1",
              "outputarea_id3",
              "user_output"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "7de5fbb0b08b40f68b29d3b8d4f7af55",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "Output()"
            ]
          },
          "metadata": {
            "tags": [
              "id3_content_1",
              "outputarea_id3",
              "user_output"
            ]
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "application/javascript": [
              "window[\"0f1105d0-f7ad-11ea-9b9e-0242ac1c0002\"] = google.colab.output.setActiveOutputArea(window[\"0f047dd8-f7ad-11ea-9b9e-0242ac1c0002\"]);\n",
              "//# sourceURL=js_9d8e291432"
            ],
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ]
          },
          "metadata": {
            "tags": [
              "id3_content_1",
              "outputarea_id3"
            ]
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "W3HS22uyAKfr"
      },
      "source": [
        "#KDE in 2D (kernel, bandwidth, sample, metric)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xPOkOFQVo_C-"
      },
      "source": [
        "Use the following demo to get the intuition on\n",
        "* How two-dimensional KDE works\n",
        "* How bandwidth influences the result\n",
        "* How choice of the kernel influences the result\n",
        "* How metric influences the result"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "npHUVDt94MAz"
      },
      "source": [
        "from sklearn.neighbors import KernelDensity\n",
        "# instantiate and fit the KDE model\n",
        "kde = KernelDensity()\n",
        "\n",
        "def fit_distribution(X, bandwidth, ker_name, metric_name):\n",
        "  params = {'bandwidth': bandwidth, 'kernel': ker_name, 'metric': metric_name}\n",
        "  kde.set_params(**params)\n",
        "  kde.fit(X)\n",
        "\n",
        "def evaluate_distribution(X):\n",
        "  logprob = kde.score_samples(X)\n",
        "  return np.exp(logprob)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R_BZreRX4P8t",
        "cellView": "form",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 617
        },
        "outputId": "b0b17250-e6c6-48d2-f8e1-c8703bd905ff"
      },
      "source": [
        "#@title Interactive 2D KDE\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from matplotlib.colors import ListedColormap\n",
        "from matplotlib import cm\n",
        "import seaborn as sns\n",
        "\n",
        "def plot_clusters(x_, y_, z_):\n",
        "  fig = plt.figure(figsize=(10,10))\n",
        "  ax = fig.gca()\n",
        "  #ax = sns.kdeplot(X[:,0], X[:,1], cmap=\"Blues\", shade=True, shade_lowest=False)\n",
        "  sns.set()\n",
        "  ax.contourf(x_, y_, z_, alpha=0.5, cmap=\"Blues\")\n",
        "  ax.tick_params(axis=\"y\",direction=\"in\", pad=-60)\n",
        "  ax.tick_params(axis=\"x\",direction=\"in\", pad=-30)\n",
        "  ax.set_xticks(np.arange(-.75, 1, .25))\n",
        "  ax.set_yticks(np.arange(-.75, 1, .25))\n",
        "  ax.set_xlim([-1, 1])\n",
        "  ax.set_ylim([-1, 1])\n",
        "\n",
        "  plt.close(fig)\n",
        "  return fig\n",
        "\n",
        "import base64\n",
        "import io\n",
        "\n",
        "def figure_to_base64(fig):\n",
        "  buf = io.BytesIO()\n",
        "  fig.savefig(buf, format='png', bbox_inches='tight', pad_inches=0);\n",
        "  image_base64 = base64.b64encode(buf.getvalue()).decode('utf-8').replace('\\n', '')\n",
        "  buf.close()\n",
        "  return image_base64\n",
        "\n",
        "\n",
        "\n",
        "import json\n",
        "from IPython.display import display, clear_output\n",
        "\n",
        "def invoke(pts, n, ker_name, metric_name):\n",
        "  data = np.array(pts).reshape((-1, 2))\n",
        "  X = data / 300.0 - 1.0\n",
        "  X[:, 1] *= -1.0\n",
        "  fit_distribution(X, float(n)/1000.0, ker_name, metric_name)\n",
        "  x_ = np.linspace(-1.0, +1.0, num=300)\n",
        "  y_ = np.linspace(-1.0, +1.0, num=300)\n",
        "  xx, yy = np.meshgrid(x_, y_)\n",
        "  x_vec = np.transpose(np.vstack([xx.ravel(), yy.ravel()]))\n",
        "  G  = evaluate_distribution(x_vec).reshape((x_.shape[0],y_.shape[0]))\n",
        "  figr = plot_clusters(x_, y_, G)\n",
        "  image_base64 = figure_to_base64(figr)\n",
        "\n",
        "  return IPython.display.JSON({'image': image_base64,\n",
        "                               'g_x': x_.tolist(), 'g_y': y_.tolist(), 'g_z': G.tolist(),\n",
        "                               's_x': X[:, 0].tolist(), 's_y': X[:, 1].tolist(), 's_z': [0]*len(X)})\n",
        "\n",
        "\n",
        "import IPython\n",
        "from google.colab import output\n",
        "\n",
        "main_str = '''\n",
        "  <script src=\"https://cdn.plot.ly/plotly-latest.min.js\"></script>\n",
        "  <script type=\"application/javascript\">\n",
        "\n",
        "    var active_pt = null;\n",
        "    var all_pts = [];\n",
        "    var classAbtn = [650, 100];\n",
        "    var classBbtn = [650, 200];\n",
        "    var ptR = 10.0;\n",
        "    var params = null;\n",
        "    var debug_txt = \"\";\n",
        "    var bg_image = new Image();\n",
        "\n",
        "    function is_close(pt1, pt2) {\n",
        "      return   (pt1[0] - pt2[0])*(pt1[0] - pt2[0])\n",
        "            +  (pt1[1] - pt2[1])*(pt1[1] - pt2[1])\n",
        "            <= ptR*ptR;\n",
        "    }\n",
        "\n",
        "    function circ(ctx, pos) {\n",
        "      ctx.beginPath();\n",
        "      ctx.arc(pos[0], pos[1], ptR, 0.0, 2.0 * Math.PI, 0);\n",
        "      ctx.fillStyle = 'rgba(255, 128, 0, 0.5)';\n",
        "      ctx.fill();\n",
        "    }\n",
        "\n",
        "\n",
        "    function draw() {\n",
        "      var canvas = document.getElementById('canvas');\n",
        "      if (canvas.getContext) {\n",
        "        var ctx = canvas.getContext('2d');\n",
        "\n",
        "        ctx.clearRect(0, 0, canvas.width, canvas.height); // cleanup before start\n",
        "        ctx.drawImage(bg_image, 0, 0, 600, 600);\n",
        "        ctx.strokeRect(0, 0, 600, 600); // field\n",
        "\n",
        "        ctx.font = '20px serif';\n",
        "        ctx.fillStyle = 'black';\n",
        "        ctx.fillText(debug_txt, 200, 200);\n",
        "\n",
        "        if (active_pt) {\n",
        "          circ(ctx, [active_pt[0], active_pt[1]]);\n",
        "        }\n",
        "        all_pts.forEach(function (item, index) {\n",
        "          circ(ctx, [item[0], item[1]]);\n",
        "        });\n",
        "\n",
        "      }\n",
        "    }\n",
        "\n",
        "    var timer\n",
        "\n",
        "    function get_state() {\n",
        "      var rates = document.getElementsByName('rate');\n",
        "      for(var i = 0; i < rates.length; i++)\n",
        "        if(rates[i].checked)\n",
        "          return rates[i].value;\n",
        "      return \"none\";\n",
        "    }\n",
        "\n",
        "\n",
        "    function mdown_handle(evt) {\n",
        "      evt.stopPropagation();\n",
        "      x = evt.offsetX;\n",
        "      y = evt.offsetY;\n",
        "      state = get_state();\n",
        "      if (state == \"add\") {\n",
        "        if (x < 600)\n",
        "            all_pts.push([x, y]);\n",
        "        return;\n",
        "      }\n",
        "\n",
        "      idx = -1;\n",
        "      all_pts.forEach(function (item, index) {\n",
        "        if (is_close([x,y], [item[0], item[1]])) {\n",
        "          idx = index;\n",
        "        }\n",
        "      });\n",
        "      if (idx <= -1)\n",
        "          return;\n",
        "      if (state == \"move\") {\n",
        "        active_pt = all_pts[idx];\n",
        "        all_pts.splice(idx, 1);\n",
        "      }\n",
        "      if (state == \"del\")\n",
        "          all_pts.splice(idx, 1);\n",
        "    }\n",
        "\n",
        "    function mmove_handle(evt) {\n",
        "      evt.stopPropagation();\n",
        "      if (get_state() != \"move\")\n",
        "          return;\n",
        "      x = evt.offsetX;\n",
        "      y = evt.offsetY;\n",
        "      if (active_pt) {\n",
        "        active_pt[0] = x;\n",
        "        active_pt[1] = y;\n",
        "      }\n",
        "    }\n",
        "\n",
        "    function mup_handle(evt) {\n",
        "      evt.stopPropagation();\n",
        "      x = evt.offsetX;\n",
        "      y = evt.offsetY;\n",
        "      if (x >= 600) {\n",
        "        active_pt = null;\n",
        "      }\n",
        "      if (active_pt) {\n",
        "        all_pts.push(active_pt);\n",
        "        active_pt = null;\n",
        "      }\n",
        "    }\n",
        "\n",
        "    function clear_field() {\n",
        "      active_pt = null;\n",
        "      all_pts = [];\n",
        "    }\n",
        "\n",
        "\n",
        "   function plot_3d(g_x, g_y, g_z)\n",
        "   {\n",
        "     cell_for_plot = document.getElementById('for_plotly_plot');\n",
        "     //Plotly.deleteTraces(cell_for_plot, [0, 1]);\n",
        "     var data = [{\n",
        "           x: g_x,\n",
        "           y: g_y,\n",
        "           z: g_z,\n",
        "           type: 'surface',\n",
        "           opacity: 0.9,\n",
        "           showscale: false,\n",
        "           colorscale: \"Viridis\"\n",
        "        }];\n",
        "\n",
        "     var layout = {\n",
        "          title: 'pdf estimate',\n",
        "          autosize: false,\n",
        "          width: 600,\n",
        "          height: 600,\n",
        "          showlegend: false,\n",
        "          scene: {\n",
        "            camera: {\n",
        "              up: {x: 0, y: 1, z: 0},\n",
        "              center: {x: 0, y: 0, z: 0},\n",
        "              eye: {x: 0.0, y: -1.0, z: 1.4}\n",
        "            }\n",
        "          }\n",
        "        };\n",
        "     Plotly.newPlot(cell_for_plot, data, layout);\n",
        "   }\n",
        "\n",
        "   function plot_scatter(s_x, s_y, s_z)\n",
        "   {\n",
        "\n",
        "     cell_for_plot = document.getElementById('for_plotly_plot');\n",
        "     var data = [{\n",
        "           x: s_x,\n",
        "           y: s_y,\n",
        "           z: s_z,\n",
        "           type: 'scatter3d',\n",
        "           mode: 'markers',\n",
        "           marker: {\n",
        "             size: 4,\n",
        "             colorscale: \"Viridis\",\n",
        "             line: {\n",
        "               width: 2,\n",
        "               color: 'rgba(0, 0, 0, 0.5)'\n",
        "             }\n",
        "           }\n",
        "        }];\n",
        "    Plotly.plot(cell_for_plot, data, {});\n",
        "   }\n",
        "\n",
        "\n",
        "    var cur_depth = 100;\n",
        "\n",
        "    async function calculate() {\n",
        "      if (!all_pts)\n",
        "        return;\n",
        "      var merged = [].concat.apply([], all_pts);\n",
        "      var e = document.getElementById(\"select_kernel\");\n",
        "      var ker_name = e.options[e.selectedIndex].value;\n",
        "      var e1 = document.getElementById(\"select_metric\");\n",
        "      var metric_name = e1.options[e1.selectedIndex].value;\n",
        "      const result = await google.colab.kernel.invokeFunction('notebook.InvSVM', [merged, cur_depth, ker_name, metric_name], {});\n",
        "      params = result.data['application/json'];\n",
        "\n",
        "      bg_image.src = 'data:image/  png;base64,' + params.image;\n",
        "      cell_for_plot = document.getElementById('for_plotly_plot');\n",
        "\n",
        "      plot_3d(params.g_x, params.g_y, params.g_z);\n",
        "      plot_scatter(params.s_x, params.s_y, params.s_z);\n",
        "\n",
        "    }\n",
        "\n",
        "   function h_change(x) {\n",
        "    el = document.getElementById(\"slider_val\");\n",
        "    cur_depth = x;\n",
        "    el.innerHTML = x;\n",
        "  }\n",
        "\n",
        " </script>\n",
        "\n",
        "<div id=\"everything\" style=\"height:600px;\">\n",
        "<div id=\"for_plotly_plot\" style=\"position:absolute; top:0px; left:0px; height:600px; width:600px;\">\n",
        "</div>\n",
        "<div id=\"custom_controls\" style=\"position:absolute; top:0px; left:600px;\">\n",
        "  <canvas id=\"canvas\" width=\"800\" height=\"600\"\n",
        "    onmousedown=\"mdown_handle(event)\"\n",
        "    onmousemove=\"mmove_handle(event)\"\n",
        "    onmouseup=\"mup_handle(event)\"></canvas>\n",
        "  <div style=\"position:absolute; top:200px; left:625px;\">\n",
        "   <div style=\"border-style: solid;\">\n",
        "    <input type=\"radio\" id=\"r1\" name=\"rate\" value=\"add\" checked=\"checked\"> Add data point\n",
        "    <br>\n",
        "    <input type=\"radio\" id=\"r2\" name=\"rate\" value=\"move\"> Move data point\n",
        "    <br>\n",
        "    <input type=\"radio\" id=\"r3\" name=\"rate\" value=\"del\"> Delete data point\n",
        "    <br>\n",
        "    <button type=\"button\" onclick=\"clear_field();\">\n",
        "      Clear\n",
        "    </button>\n",
        "   </div>\n",
        "   <br><br>Bandwidth x 1000:<br>\n",
        "    <input type=\"range\" min=\"1\" max=\"200\" value=\"100\" oninput=\"h_change(this.value)\">\n",
        "    <br><div id=\"slider_val\">100</div>\n",
        "   <br>Kernel:<br>\n",
        "   <select id=\"select_kernel\">\n",
        "       <option value=\"gaussian\">gaussian</option>\n",
        "       <option value=\"tophat\">tophat</option>\n",
        "       <option value=\"epanechnikov\">epanechnikov</option>\n",
        "       <option value=\"exponential\">exponential</option>\n",
        "       <option value=\"linear\">linear</option>\n",
        "       <option value=\"cosine\">cosine</option>\n",
        "   </select>\n",
        "   <br>\n",
        "   Metric:\n",
        "   <br>\n",
        "   <select id=\"select_metric\">\n",
        "       <option value=\"euclidean\">euclidean</option>\n",
        "       <option value=\"manhattan\">manhattan</option>\n",
        "       <option value=\"infinity\">infinity</option>\n",
        "   </select>\n",
        "   <br><br>\n",
        "    <button type=\"button\" onclick=\"calculate();\">\n",
        "      KDE\n",
        "    </button>\n",
        "  </div>\n",
        "</div>\n",
        "<div>\n",
        "\n",
        "<script>\n",
        "function init()\n",
        "{\n",
        "  timer = setInterval(draw, 10);\n",
        "}\n",
        "\n",
        "init();\n",
        "</script>\n",
        "'''\n",
        "\n",
        "display(IPython.display.HTML(main_str))\n",
        "output.register_callback('notebook.InvSVM', invoke)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "  <script src=\"https://cdn.plot.ly/plotly-latest.min.js\"></script> \n",
              "  <script type=\"application/javascript\">\n",
              "  \n",
              "    var active_pt = null;\n",
              "    var all_pts = [];\n",
              "    var classAbtn = [650, 100];\n",
              "    var classBbtn = [650, 200];\n",
              "    var ptR = 10.0;\n",
              "    var params = null;\n",
              "    var debug_txt = \"\";\n",
              "    var bg_image = new Image();\n",
              "\n",
              "    function is_close(pt1, pt2) {\n",
              "      return   (pt1[0] - pt2[0])*(pt1[0] - pt2[0])\n",
              "            +  (pt1[1] - pt2[1])*(pt1[1] - pt2[1])\n",
              "            <= ptR*ptR;\n",
              "    }\n",
              "  \n",
              "    function circ(ctx, pos) {\n",
              "      ctx.beginPath();\n",
              "      ctx.arc(pos[0], pos[1], ptR, 0.0, 2.0 * Math.PI, 0);\n",
              "      ctx.fillStyle = 'rgba(255, 128, 0, 0.5)';\n",
              "      ctx.fill();\n",
              "    }\n",
              "    \n",
              "    \n",
              "    function draw() {\n",
              "      var canvas = document.getElementById('canvas');\n",
              "      if (canvas.getContext) {\n",
              "        var ctx = canvas.getContext('2d');\n",
              "        \n",
              "        ctx.clearRect(0, 0, canvas.width, canvas.height); // cleanup before start\n",
              "        ctx.drawImage(bg_image, 0, 0, 600, 600);\n",
              "        ctx.strokeRect(0, 0, 600, 600); // field\n",
              "        \n",
              "        ctx.font = '20px serif';\n",
              "        ctx.fillStyle = 'black';\n",
              "        ctx.fillText(debug_txt, 200, 200);\n",
              "        \n",
              "        if (active_pt) {\n",
              "          circ(ctx, [active_pt[0], active_pt[1]]);\n",
              "        }\n",
              "        all_pts.forEach(function (item, index) {\n",
              "          circ(ctx, [item[0], item[1]]);\n",
              "        });\n",
              "\n",
              "      }\n",
              "    }\n",
              "    \n",
              "    var timer\n",
              "\n",
              "    function get_state() {\n",
              "      var rates = document.getElementsByName('rate');\n",
              "      for(var i = 0; i < rates.length; i++)\n",
              "        if(rates[i].checked)\n",
              "          return rates[i].value;\n",
              "      return \"none\";\n",
              "    }\n",
              "\n",
              "    \n",
              "    function mdown_handle(evt) {\n",
              "      evt.stopPropagation();\n",
              "      x = evt.offsetX;\n",
              "      y = evt.offsetY;\n",
              "      state = get_state();\n",
              "      if (state == \"add\") {\n",
              "        if (x < 600)\n",
              "            all_pts.push([x, y]);\n",
              "        return;\n",
              "      }\n",
              "\n",
              "      idx = -1;\n",
              "      all_pts.forEach(function (item, index) {\n",
              "        if (is_close([x,y], [item[0], item[1]])) {\n",
              "          idx = index;\n",
              "        }\n",
              "      });\n",
              "      if (idx <= -1)\n",
              "          return;\n",
              "      if (state == \"move\") {\n",
              "        active_pt = all_pts[idx];\n",
              "        all_pts.splice(idx, 1);\n",
              "      }\n",
              "      if (state == \"del\")\n",
              "          all_pts.splice(idx, 1);\n",
              "    }\n",
              "    \n",
              "    function mmove_handle(evt) {\n",
              "      evt.stopPropagation();\n",
              "      if (get_state() != \"move\")\n",
              "          return;\n",
              "      x = evt.offsetX;\n",
              "      y = evt.offsetY;\n",
              "      if (active_pt) {\n",
              "        active_pt[0] = x;\n",
              "        active_pt[1] = y;\n",
              "      }\n",
              "    }\n",
              "    \n",
              "    function mup_handle(evt) {\n",
              "      evt.stopPropagation();\n",
              "      x = evt.offsetX;\n",
              "      y = evt.offsetY;\n",
              "      if (x >= 600) {\n",
              "        active_pt = null;\n",
              "      }\n",
              "      if (active_pt) {\n",
              "        all_pts.push(active_pt);\n",
              "        active_pt = null;\n",
              "      }\n",
              "    }\n",
              "    \n",
              "    function clear_field() {\n",
              "      active_pt = null;\n",
              "      all_pts = [];\n",
              "    }\n",
              "\n",
              "\n",
              "   function plot_3d(g_x, g_y, g_z)\n",
              "   {\n",
              "     cell_for_plot = document.getElementById('for_plotly_plot');\n",
              "     //Plotly.deleteTraces(cell_for_plot, [0, 1]);\n",
              "     var data = [{\n",
              "           x: g_x,\n",
              "           y: g_y,\n",
              "           z: g_z,\n",
              "           type: 'surface',\n",
              "           opacity: 0.9,\n",
              "           showscale: false,\n",
              "           colorscale: \"Viridis\"\n",
              "        }];\n",
              "  \n",
              "     var layout = {\n",
              "          title: 'pdf estimate',\n",
              "          autosize: false,\n",
              "          width: 600,\n",
              "          height: 600,\n",
              "          showlegend: false,\n",
              "          scene: {\n",
              "            camera: {\n",
              "              up: {x: 0, y: 1, z: 0},\n",
              "              center: {x: 0, y: 0, z: 0},\n",
              "              eye: {x: 0.0, y: -1.0, z: 1.4}\n",
              "            }\n",
              "          }\n",
              "        };\n",
              "     Plotly.newPlot(cell_for_plot, data, layout);\n",
              "   }\n",
              "\n",
              "   function plot_scatter(s_x, s_y, s_z)\n",
              "   {\n",
              "          \n",
              "     cell_for_plot = document.getElementById('for_plotly_plot');\n",
              "     var data = [{\n",
              "           x: s_x,\n",
              "           y: s_y,\n",
              "           z: s_z,\n",
              "           type: 'scatter3d',\n",
              "           mode: 'markers',\n",
              "           marker: {\n",
              "             size: 4,\n",
              "             colorscale: \"Viridis\",\n",
              "             line: {\n",
              "               width: 2,\n",
              "               color: 'rgba(0, 0, 0, 0.5)'\n",
              "             }\n",
              "           }\n",
              "        }];\n",
              "    Plotly.plot(cell_for_plot, data, {});\n",
              "   }\n",
              "\n",
              "\n",
              "    var cur_depth = 100;\n",
              "\n",
              "    async function calculate() {\n",
              "      if (!all_pts)\n",
              "        return;\n",
              "      var merged = [].concat.apply([], all_pts);\n",
              "      var e = document.getElementById(\"select_kernel\");\n",
              "      var ker_name = e.options[e.selectedIndex].value;\n",
              "      var e1 = document.getElementById(\"select_metric\");\n",
              "      var metric_name = e1.options[e1.selectedIndex].value;\n",
              "      const result = await google.colab.kernel.invokeFunction('notebook.InvSVM', [merged, cur_depth, ker_name, metric_name], {});\n",
              "      params = result.data['application/json'];\n",
              "            \n",
              "      bg_image.src = 'data:image/  png;base64,' + params.image;\n",
              "      cell_for_plot = document.getElementById('for_plotly_plot');\n",
              "      \n",
              "      plot_3d(params.g_x, params.g_y, params.g_z);\n",
              "      plot_scatter(params.s_x, params.s_y, params.s_z);\n",
              "      \n",
              "    }\n",
              "\n",
              "   function h_change(x) {\n",
              "    el = document.getElementById(\"slider_val\");\n",
              "    cur_depth = x;\n",
              "    el.innerHTML = x;\n",
              "  }\n",
              "\n",
              " </script>\n",
              "\n",
              "<div id=\"everything\" style=\"height:600px;\">\n",
              "<div id=\"for_plotly_plot\" style=\"position:absolute; top:0px; left:0px; height:600px; width:600px;\">\n",
              "</div>\n",
              "<div id=\"custom_controls\" style=\"position:absolute; top:0px; left:600px;\">\n",
              "  <canvas id=\"canvas\" width=\"800\" height=\"600\" \n",
              "    onmousedown=\"mdown_handle(event)\"\n",
              "    onmousemove=\"mmove_handle(event)\"\n",
              "    onmouseup=\"mup_handle(event)\"></canvas>\n",
              "  <div style=\"position:absolute; top:200px; left:625px;\">\n",
              "   <div style=\"border-style: solid;\">\n",
              "    <input type=\"radio\" id=\"r1\" name=\"rate\" value=\"add\" checked=\"checked\"> Add data point\n",
              "    <br>\n",
              "    <input type=\"radio\" id=\"r2\" name=\"rate\" value=\"move\"> Move data point\n",
              "    <br>\n",
              "    <input type=\"radio\" id=\"r3\" name=\"rate\" value=\"del\"> Delete data point\n",
              "    <br>\n",
              "    <button type=\"button\" onclick=\"clear_field();\">\n",
              "      Clear\n",
              "    </button>\n",
              "   </div>\n",
              "   <br><br>Bandwidth x 1000:<br>\n",
              "    <input type=\"range\" min=\"1\" max=\"200\" value=\"100\" oninput=\"h_change(this.value)\">\n",
              "    <br><div id=\"slider_val\">100</div>\n",
              "   <br>Kernel:<br>\n",
              "   <select id=\"select_kernel\">\n",
              "       <option value=\"gaussian\">gaussian</option>\n",
              "       <option value=\"tophat\">tophat</option>\n",
              "       <option value=\"epanechnikov\">epanechnikov</option>\n",
              "       <option value=\"exponential\">exponential</option>\n",
              "       <option value=\"linear\">linear</option>\n",
              "       <option value=\"cosine\">cosine</option>\n",
              "   </select>\n",
              "   <br>\n",
              "   Metric:\n",
              "   <br>\n",
              "   <select id=\"select_metric\">\n",
              "       <option value=\"euclidean\">euclidean</option>\n",
              "       <option value=\"manhattan\">manhattan</option>\n",
              "       <option value=\"infinity\">infinity</option>\n",
              "   </select>\n",
              "   <br><br>\n",
              "    <button type=\"button\" onclick=\"calculate();\">\n",
              "      KDE\n",
              "    </button>\n",
              "  </div>\n",
              "</div>\n",
              "<div>\n",
              "  \n",
              "<script>\n",
              "function init()\n",
              "{\n",
              "  timer = setInterval(draw, 10);\n",
              "}\n",
              "\n",
              "init();\n",
              "</script>\n"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6RG1GigA-C6r"
      },
      "source": [
        "#1849 Broad Street cholera outbreak visualization"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jJ3VwZ_fMwvo"
      },
      "source": [
        "The Broad Street cholera outbreak was a severe outbreak of cholera that occurred in 1854 near Broad Street (now Broadwick Street) in the Soho district of the City of Westminster, London, England, and occurred during the 1846–1860 cholera pandemic happening worldwide. This outbreak, which killed 616 people, is best known for the physician John Snow's study of its causes and his hypothesis that germ-contaminated water was the source of cholera, rather than particles in the air.This discovery came to influence public health and the construction of improved sanitation facilities beginning in the mid-19th century."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4fXVGgGf-F5T"
      },
      "source": [
        "\n",
        "![](https://raw.githubusercontent.com/fbeilstein/machine_learning/master/lecture_15_kernel_density_estimation/snows.png)\n",
        "\n",
        "On 31 August 1854, after several other outbreaks had occurred elsewhere in the city, a major outbreak of cholera occurred in Soho. Snow, the physician who eventually linked the outbreak to contaminated water, later called it \"the most terrible outbreak of cholera which ever occurred in this kingdom.\"\n",
        "\n",
        "Over the next three days, 127 people on or near Broad Street died. During the next week, three quarters of the residents had fled the area. By 10 September, 500 people had died and the mortality rate was 12.8 percent in some parts of the city. By the end of the outbreak, 616 people had died.\n",
        "\n",
        "Many of the victims were taken to the Middlesex Hospital, where their treatment was superintended by Florence Nightingale (we've talked about here in statistics lecture), who briefly joined the hospital in early September in order to help with the outbreak.\n",
        "\n",
        "By talking to local residents Snow identified the source of the outbreak as the public water pump on Broad Street. Although Snow's chemical and microscope examination of a sample of the water from this Broad Street pump water did not conclusively prove its danger, his facts about the patterns of illness and death among residents in Soho persuaded the St James parish authorities to disable the well pump by removing its handle.\n",
        "\n",
        "It's worth mentioning that the germ theory was not established at this point (Louis Pasteur did not propose it until 1861). Snow did not understand the mechanism by which disease was transmitted, but the evidence led him to believe that it was NOT due to breathing foul air. Based on the pattern of illness among residents, Snow hypothesized that cholera was spread by an agent in contaminated water. He first published his theory in 1849, in an essay titled \"On the Mode of Communication of Cholera\".\n",
        "\n",
        "<br>\n",
        "\n",
        "<img src=\"https://raw.githubusercontent.com/fbeilstein/machine_learning/master/lecture_15_kernel_density_estimation/snows_cholera_map.png\" width=\"650px\" />\n",
        "\n",
        ""
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fSfWfk_LHhbe"
      },
      "source": [
        "**Preparation**\n",
        "\n",
        "Download the dataset in a `csv` format from github.\n",
        "Following code should do the job."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B1KTWDE7Hde5"
      },
      "source": [
        "# download the dataset\n",
        "!wget https://raw.githubusercontent.com/fbeilstein/machine_learning/master/datasets/pumps.csv\n",
        "!wget https://raw.githubusercontent.com/fbeilstein/machine_learning/master/datasets/deaths.csv\n",
        "!wget https://raw.githubusercontent.com/fbeilstein/machine_learning/master/datasets/cholera_map.jpg"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-OLCG1TUTkFl"
      },
      "source": [
        "The files you have downloaded contain the following information\n",
        "* `pumps.csv` -- coordinates of pumps (sources of water), contains $2$ columns, namely `x` and `y`\n",
        "* `deaths.csv` -- coordinates of households, where anyone had died from cholera. Points can repeat if few deaths in one house were spotted.\n",
        "* `cholera_map.jpg` -- map of London. Coordinates in other files correspond to pixels in this image."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FOSvo-Njcnjc"
      },
      "source": [
        "**Problem 1:**\n",
        "complete the `read_dataframe(filename)` function so that it takes name of the file and returns pandas.dataframe"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "J2akMqyzaDwz"
      },
      "source": [
        "import pandas as pd\n",
        "\n",
        "def read_dataframe(filename):\n",
        "  # ENTER YOUR CODE HERE"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5pZNH6Prc0xp"
      },
      "source": [
        "Use the following code to visualize tables."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1LHzjTJGbggm",
        "cellView": "form"
      },
      "source": [
        "#@title Visualize table\n",
        "from google.colab import data_table\n",
        "from google.colab import widgets\n",
        "\n",
        "tb = widgets.TabBar(['pumps.csv', 'deaths.csv'], location='top')\n",
        "with tb.output_to(0):\n",
        "  display(data_table.DataTable(read_dataframe('pumps.csv'), include_index=False, num_rows_per_page=10))\n",
        "with tb.output_to(1):\n",
        "  display(data_table.DataTable(read_dataframe('deaths.csv'), include_index=False, num_rows_per_page=10))\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "F_Hkog6_edhH"
      },
      "source": [
        "**Problem 2:**\n",
        "Visualize `cholera_map.jpg` with matplotlib.\n",
        "Complete the following function so that it plots the map.\n",
        "You can use `imread` and `imshow` functions."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u0s685vlfDuL"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "def plot_map(filename):\n",
        "  # ENTER YOUR CODE HERE\n",
        "\n",
        "plt.figure(figsize=(15, 15));\n",
        "plot_map('cholera_map.jpg')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "exmVtE0LjbuP"
      },
      "source": [
        "**Problem 3:**\n",
        "Visualize points from given dataframe as a scatter plot.\n",
        "Suppose that dataframe contains two columns `x` and `y`.\n",
        "Use parameter `color` to set the color of points."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "meUNoFQ0jqDj"
      },
      "source": [
        "def plot_points(dataframe, color):\n",
        "  # ENTER YOUR CODE HERE\n",
        "\n",
        "plt.figure(figsize=(15, 15));\n",
        "plot_map('cholera_map.jpg')\n",
        "plot_points(read_dataframe('deaths.csv'), 'b')\n",
        "plot_points(read_dataframe('pumps.csv'), 'r')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_nEyVZ72k1Y2"
      },
      "source": [
        "**Problem 4:**\n",
        "Visualize points from the dataframe given as a kernel density.\n",
        "For better visuals tune opacity (e.g. set `alpha=0.8`).\n",
        "You may suppose that dataframe contains columns `x` and `y`."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eIVZAg-SlJ8D"
      },
      "source": [
        "import seaborn as sns\n",
        "\n",
        "def plot_kde(dataframe):\n",
        "  # ENTER YOUR CODE HERE\n",
        "\n",
        "plt.figure(figsize=(15, 15));\n",
        "plot_map('cholera_map.jpg')\n",
        "plot_kde(read_dataframe('deaths.csv'))\n",
        "plot_points(read_dataframe('pumps.csv'), 'r')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gmtCPkI9lv75",
        "cellView": "form"
      },
      "source": [
        "#@title Visualize both\n",
        "from google.colab import widgets\n",
        "\n",
        "tb = widgets.TabBar(['Scatter', 'KDE'], location='top')\n",
        "with tb.output_to(0):\n",
        "  plt.figure(figsize=(15, 15));\n",
        "  plot_map('cholera_map.jpg')\n",
        "  plot_points(read_dataframe('deaths.csv'), 'b')\n",
        "  plot_points(read_dataframe('pumps.csv'), 'r')\n",
        "  plt.ylabel('y')\n",
        "  plt.xlabel('x')\n",
        "with tb.output_to(1):\n",
        "  plt.figure(figsize=(15, 15));\n",
        "  plot_map('cholera_map.jpg')\n",
        "  plot_kde(read_dataframe('deaths.csv'))\n",
        "  plot_points(read_dataframe('pumps.csv'), 'r')\n",
        "  plt.ylabel('y')\n",
        "  plt.xlabel('x')"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}