Sentiment_Analysis_Multi_label_Classification_on_E_Commerce_Reviews.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.6"
    },
    "colab": {
      "name": "Sentiment Analysis Multi-label Classification on E-Commerce Reviews.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "93dd867f6c424abf986a39e6d16aa389": {
          "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_92b219d31d89435da2793956431c0215",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_e6df8030d3ca4e85a4d197d7c1268d78",
              "IPY_MODEL_2c7c06e8e7374c33a3433765dad28667"
            ]
          }
        },
        "92b219d31d89435da2793956431c0215": {
          "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
          }
        },
        "e6df8030d3ca4e85a4d197d7c1268d78": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_9191dad4453544388092bdac04522114",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 433,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 433,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_0b3ca3afad23462180cc47208227d0f9"
          }
        },
        "2c7c06e8e7374c33a3433765dad28667": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_d3da291fcd784edea6b3db29c799cfa0",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 433/433 [00:22&lt;00:00, 19.1B/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_4ce78e8e3eea45b5b2b815515f64afd7"
          }
        },
        "9191dad4453544388092bdac04522114": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "0b3ca3afad23462180cc47208227d0f9": {
          "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
          }
        },
        "d3da291fcd784edea6b3db29c799cfa0": {
          "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"
          }
        },
        "4ce78e8e3eea45b5b2b815515f64afd7": {
          "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
          }
        },
        "de2f20f766bd40b39f8c3cce077d1207": {
          "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_528e15a242e140a7aea4377f3a8860e9",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_d146333b63624bfe98163ad89bdcaaf3",
              "IPY_MODEL_0f340f2eca6749188812e2f212b94b1b"
            ]
          }
        },
        "528e15a242e140a7aea4377f3a8860e9": {
          "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
          }
        },
        "d146333b63624bfe98163ad89bdcaaf3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_2f3ada5798734d73b3b99306c14d28d0",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 536063208,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 536063208,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_6eb0d451d694459194022a982c504dc3"
          }
        },
        "0f340f2eca6749188812e2f212b94b1b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_6478c92973bb4e0d933af010e865ef59",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 536M/536M [00:21&lt;00:00, 24.8MB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_b1dd7340424c4611aeefc485d5ee5358"
          }
        },
        "2f3ada5798734d73b3b99306c14d28d0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "6eb0d451d694459194022a982c504dc3": {
          "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
          }
        },
        "6478c92973bb4e0d933af010e865ef59": {
          "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"
          }
        },
        "b1dd7340424c4611aeefc485d5ee5358": {
          "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
          }
        },
        "9320ce02c1684ef29557b7ce9bae9e7d": {
          "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_94d6964281914b029c70996a727c5092",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_08f4f7a1d20f420eaf8289c3ca4ddcac",
              "IPY_MODEL_eeb1247c1db2481daff83dc07d98a949"
            ]
          }
        },
        "94d6964281914b029c70996a727c5092": {
          "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
          }
        },
        "08f4f7a1d20f420eaf8289c3ca4ddcac": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_a1148332a53e4561ad7048d549a77599",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 231508,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 231508,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_8a48d78aa20540998b3673b6255d55da"
          }
        },
        "eeb1247c1db2481daff83dc07d98a949": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_c19805df136b40c286903770668c3742",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 232k/232k [00:00&lt;00:00, 251kB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_dae8d41fb7d2486c87083726a34a8ede"
          }
        },
        "a1148332a53e4561ad7048d549a77599": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "8a48d78aa20540998b3673b6255d55da": {
          "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
          }
        },
        "c19805df136b40c286903770668c3742": {
          "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"
          }
        },
        "dae8d41fb7d2486c87083726a34a8ede": {
          "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
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/YanisNC/Sentiment-Analysis-Multi-Label-Classification-E-Commerce-Reviews/blob/main/Sentiment_Analysis_Multi_label_Classification_on_E_Commerce_Reviews.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9PanEhGvYgXn"
      },
      "source": [
        "# Sentiment Analysis Multi-label Classification on E-commerce Reviews\r\n",
        "\r\n",
        "- The dataset contains 10000 samples of reviews and ratings about a fashion e-commerce.\r\n",
        "\r\n",
        "- The reviews have been preprocessed using NLTK.\r\n",
        "\r\n",
        "- BERT has been used for fine-tuning training."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dpvAJLy8XesN"
      },
      "source": [
        "### Import libraries"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0feK_AQrd1QM"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import pandas as pd\n",
        "import tensorflow as tf\n",
        "import nltk\n",
        "from nltk.corpus import stopwords\n",
        "import string"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZOXTUDSEvxyL",
        "outputId": "2b4e508c-c48a-4099-cd30-f80461cd91c1"
      },
      "source": [
        "nltk.download('stopwords')\r\n",
        "nltk.download('wordnet')\r\n",
        "nltk.download('punkt')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[nltk_data] Downloading package stopwords to /root/nltk_data...\n",
            "[nltk_data]   Package stopwords is already up-to-date!\n",
            "[nltk_data] Downloading package wordnet to /root/nltk_data...\n",
            "[nltk_data]   Package wordnet is already up-to-date!\n",
            "[nltk_data] Downloading package punkt to /root/nltk_data...\n",
            "[nltk_data]   Unzipping tokenizers/punkt.zip.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JDcODDUAeBEs",
        "outputId": "43ec12d4-2faa-41f5-a81b-0ec2b19af612"
      },
      "source": [
        "pip install transformers"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting transformers\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/f9/54/5ca07ec9569d2f232f3166de5457b63943882f7950ddfcc887732fc7fb23/transformers-4.3.3-py3-none-any.whl (1.9MB)\n",
            "\u001b[K     |████████████████████████████████| 1.9MB 11.3MB/s \n",
            "\u001b[?25hRequirement already satisfied: numpy>=1.17 in /usr/local/lib/python3.7/dist-packages (from transformers) (1.19.5)\n",
            "Collecting sacremoses\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/7d/34/09d19aff26edcc8eb2a01bed8e98f13a1537005d31e95233fd48216eed10/sacremoses-0.0.43.tar.gz (883kB)\n",
            "\u001b[K     |████████████████████████████████| 890kB 54.3MB/s \n",
            "\u001b[?25hRequirement already satisfied: packaging in /usr/local/lib/python3.7/dist-packages (from transformers) (20.9)\n",
            "Requirement already satisfied: tqdm>=4.27 in /usr/local/lib/python3.7/dist-packages (from transformers) (4.41.1)\n",
            "Collecting tokenizers<0.11,>=0.10.1\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/71/23/2ddc317b2121117bf34dd00f5b0de194158f2a44ee2bf5e47c7166878a97/tokenizers-0.10.1-cp37-cp37m-manylinux2010_x86_64.whl (3.2MB)\n",
            "\u001b[K     |████████████████████████████████| 3.2MB 36.1MB/s \n",
            "\u001b[?25hRequirement already satisfied: importlib-metadata; python_version < \"3.8\" in /usr/local/lib/python3.7/dist-packages (from transformers) (3.7.0)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.7/dist-packages (from transformers) (3.0.12)\n",
            "Requirement already satisfied: regex!=2019.12.17 in /usr/local/lib/python3.7/dist-packages (from transformers) (2019.12.20)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.7/dist-packages (from transformers) (2.23.0)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.7/dist-packages (from sacremoses->transformers) (1.15.0)\n",
            "Requirement already satisfied: click in /usr/local/lib/python3.7/dist-packages (from sacremoses->transformers) (7.1.2)\n",
            "Requirement already satisfied: joblib in /usr/local/lib/python3.7/dist-packages (from sacremoses->transformers) (1.0.1)\n",
            "Requirement already satisfied: pyparsing>=2.0.2 in /usr/local/lib/python3.7/dist-packages (from packaging->transformers) (2.4.7)\n",
            "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.7/dist-packages (from importlib-metadata; python_version < \"3.8\"->transformers) (3.4.0)\n",
            "Requirement already satisfied: typing-extensions>=3.6.4; python_version < \"3.8\" in /usr/local/lib/python3.7/dist-packages (from importlib-metadata; python_version < \"3.8\"->transformers) (3.7.4.3)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests->transformers) (2020.12.5)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests->transformers) (2.10)\n",
            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests->transformers) (3.0.4)\n",
            "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.7/dist-packages (from requests->transformers) (1.24.3)\n",
            "Building wheels for collected packages: sacremoses\n",
            "  Building wheel for sacremoses (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for sacremoses: filename=sacremoses-0.0.43-cp37-none-any.whl size=893262 sha256=23854776a4fe8e2272ffd8deac6c72bc9e165d914acf28689d8aefa984328532\n",
            "  Stored in directory: /root/.cache/pip/wheels/29/3c/fd/7ce5c3f0666dab31a50123635e6fb5e19ceb42ce38d4e58f45\n",
            "Successfully built sacremoses\n",
            "Installing collected packages: sacremoses, tokenizers, transformers\n",
            "Successfully installed sacremoses-0.0.43 tokenizers-0.10.1 transformers-4.3.3\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nt4bayVXXizJ"
      },
      "source": [
        "### Import dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1SLzDrm9d1QT"
      },
      "source": [
        "df = pd.read_csv(\"ratings.csv\", sep=',')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "scrolled": true,
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "poAfUC5zd1QU",
        "outputId": "2fc2fbc9-f929-4d80-b601-6f08bf9293bc"
      },
      "source": [
        "df.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>TEXT</th>\n",
              "      <th>RATING</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>The shirt was more of a smock. I expected a so...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>The shirt was more of a smock. I expected a so...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>The shirt was more of a smock. I expected a so...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Just received my order today. When I opened th...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>First of all, the button hole on this belt nev...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                TEXT  RATING\n",
              "0  The shirt was more of a smock. I expected a so...       1\n",
              "1  The shirt was more of a smock. I expected a so...       1\n",
              "2  The shirt was more of a smock. I expected a so...       1\n",
              "3  Just received my order today. When I opened th...       1\n",
              "4  First of all, the button hole on this belt nev...       1"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "60-QvXxcX03G"
      },
      "source": [
        "### Exploratory Data Analysis"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "C5ZnFWoHd1QV"
      },
      "source": [
        "The first three rows are identical. Let's remove duplicate rows."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "scrolled": true,
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YHg9VoBid1QW",
        "outputId": "8a7dbcee-4f07-4c7c-a2b7-73d5966a4234"
      },
      "source": [
        "df.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(10000, 2)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "scrolled": true,
        "id": "ZmggiHJ9d1QW"
      },
      "source": [
        "df = df.drop_duplicates()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AxbX5199d1QW",
        "outputId": "9a957aab-4937-428f-c77f-d1f14fa78e71"
      },
      "source": [
        "df.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(3467, 2)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UJ_HzywL4w5_",
        "outputId": "c092f76f-95df-4083-c5f2-bdcc32699c69"
      },
      "source": [
        "# Check if there are any NaN values.\r\n",
        "df.isna().any()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "TEXT      False\n",
              "RATING    False\n",
              "dtype: bool"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FJqwEORMd1QX"
      },
      "source": [
        "As we can see, 6533 rows are duplicated. These rows must be removed to avoid overfitting."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "8rzvvAuld1QX",
        "outputId": "e2c02033-c389-4c02-cadf-3ef20046f870"
      },
      "source": [
        "sns.countplot(x=df[\"RATING\"])\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Tywtgf6mYHgM"
      },
      "source": [
        "The classes are quite balanced"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yFFIWnpog7pR"
      },
      "source": [
        " #### Analyze the labels"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0AMXmJe1hCFZ",
        "outputId": "ad3f8ae7-4a9c-4b0b-81fe-bef4957ff899"
      },
      "source": [
        "df[\"RATING\"].unique()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([1, 2, 3, 5, 4])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "C5orJMQphXth",
        "outputId": "aae45ae5-5f56-44d1-e113-a9766267631a"
      },
      "source": [
        "# Let's change it so it starts at 0\r\n",
        "df[\"RATING\"] = df[\"RATING\"] - 1\r\n",
        "\r\n",
        "df[\"RATING\"].unique()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 1, 2, 4, 3])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "RsFS4ZaIpnhS",
        "outputId": "b1d206cd-4fe4-4e0f-c44e-0ff3838f2e3b"
      },
      "source": [
        "sns.countplot(x=df[\"RATING\"])\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gdRvwRe0w7cH"
      },
      "source": [
        "rating_1 = []\r\n",
        "for review in df.loc[df[\"RATING\"] == 0][\"TEXT\"]:\r\n",
        "  token = nltk.word_tokenize(review)\r\n",
        "  rating_1.extend(token)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 414
        },
        "id": "VLcxCXJCuRvY",
        "outputId": "75d7497e-9843-4d1c-ce45-566e26f7cbe8"
      },
      "source": [
        " #calculate word frequency\r\n",
        " plt.figure(figsize=(10,6))\r\n",
        "\r\n",
        "freq = nltk.FreqDist(rating_1)\r\n",
        "freq.plot(5, cumulative=False)\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<FreqDist with 3267 samples and 31599 outcomes>\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "C-6yYHTC3haP"
      },
      "source": [
        "### Data Preprocessing"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H9UJCZJ8XeW6"
      },
      "source": [
        "<img src=https://miro.medium.com/max/700/0*plzLfemOxB1ftrKJ width=\"600\">"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nJGG1aKU_vfz"
      },
      "source": [
        "##### An example for each sentiment"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KrxHHIBk-v3l",
        "outputId": "a5a13728-830c-4f13-f5f4-8e32227f4335"
      },
      "source": [
        "print(\"Rating 1 :\",df[df['RATING']==0]['TEXT'].values[0])\r\n",
        "print(\"Rating 2 :\",df[df['RATING']==1]['TEXT'].values[0])\r\n",
        "print(\"Rating 3 :\",df[df['RATING']==2]['TEXT'].values[0])\r\n",
        "print(\"Rating 4 :\",df[df['RATING']==3]['TEXT'].values[0])\r\n",
        "print(\"Rating 5 :\",df[df['RATING']==4]['TEXT'].values[0])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Rating 1 : The shirt was more of a smock. I expected a soft little cotton t shirt, but I suppose it\\'s my fault for not reading the description carefully. Also, the white shirt was completely see through. I returned mine.\"}},{r:{id:46830320,si:1,pi:10860142,mu:\"{f7dedd4c-02d2-49ba-9553-4f9be0bdb9e5}\n",
            "Rating 2 : I liked the shirt. it\\'s pretty. but I couldn\\'t fit my arms in the shirt. the sleeves are too tight. get a size bigger if you don\\'t have twig arms\"}},{r:{id:47933895,si:1,v:3,t:3,r:1,h:\"Not what I expected\n",
            "Rating 3 : I was really excited to see this in mail today, but got very disappointed right after trying it. I'm a petite 5 ft. 3\\\n",
            "Rating 4 : I really like this dress! It\\'s so comfortable, the colors are great for the summer, and it flows so nicely! I wore it to work &amp; to a night lounge dressed up with wedges! I got many compliments! \n",
            "Rating 5 : This dress is absolutely wonderful. Was afraid it might be too long for me, but the length is perfect! Dont even need to wear heels with it. Im 5\\'4 and i got it in S.\"}},{r:{id:74223533,si:1,pi:16960075,dp:true,mu:\"maharaibsyed013@gmail.com\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WF1iHlpFvYUe"
      },
      "source": [
        "def text_preprocessing(row):\r\n",
        "\r\n",
        "  #lowercase\r\n",
        "  row = row.lower()\r\n",
        "\r\n",
        "  #stopwords\r\n",
        "  stop_words = set(stopwords.words('english'))\r\n",
        "  row = ' '.join([word for word in row.split() if word not in (stop_words)])\r\n",
        "\r\n",
        "  #punctuation\r\n",
        "  punctuation = set(string.punctuation)\r\n",
        "  row = ''.join(ch for ch in row if ch not in punctuation)\r\n",
        "\r\n",
        "  #lemmatization\r\n",
        "  wn = nltk.WordNetLemmatizer()\r\n",
        "  row = ' '.join([wn.lemmatize(word) for word in row.split()])\r\n",
        "\r\n",
        "  return row"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ONnWgNnm3IPN"
      },
      "source": [
        "df_clean = df.copy()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-_gvfTF3vcK2"
      },
      "source": [
        "df_clean['TEXT'] = df_clean['TEXT'].apply(str).apply(lambda x: text_preprocessing(x))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "muuAkjG5vwC6",
        "outputId": "4ed32883-25d8-4a01-9b96-d84325cb1564"
      },
      "source": [
        "df_clean.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>TEXT</th>\n",
              "      <th>RATING</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>shirt smock expected soft little cotton shirt ...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>received order today opened package appeared u...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>first all button hole belt never matched size ...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>didnt hold anything obvious</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>nothing else needed say</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                TEXT  RATING\n",
              "0  shirt smock expected soft little cotton shirt ...       0\n",
              "3  received order today opened package appeared u...       0\n",
              "4  first all button hole belt never matched size ...       0\n",
              "5                        didnt hold anything obvious       0\n",
              "6                            nothing else needed say       0"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 103
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "PoUEvZuv6Np-",
        "outputId": "0ffe0c73-bf7d-4157-cad4-c028ec9ae3dc"
      },
      "source": [
        "df.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>TEXT</th>\n",
              "      <th>RATING</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>The shirt was more of a smock. I expected a so...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Just received my order today. When I opened th...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>First of all, the button hole on this belt nev...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>these didnt hold anything up and was obvious i...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>Nothing else needed to say</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                TEXT  RATING\n",
              "0  The shirt was more of a smock. I expected a so...       0\n",
              "3  Just received my order today. When I opened th...       0\n",
              "4  First of all, the button hole on this belt nev...       0\n",
              "5  these didnt hold anything up and was obvious i...       0\n",
              "6                         Nothing else needed to say       0"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 104
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BghX-TZTYzII"
      },
      "source": [
        "Comparing the text before and after preprocessing"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ISHl8cgLCOzN",
        "outputId": "c3d5aa99-6457-479a-bb63-9cace0d91e39"
      },
      "source": [
        "print(\"Rating 1 :\",df[df['RATING']==0]['TEXT'].values[0])\r\n",
        "print(\"Rating 2 :\",df[df['RATING']==1]['TEXT'].values[0])\r\n",
        "print(\"Rating 3 :\",df[df['RATING']==2]['TEXT'].values[0])\r\n",
        "print(\"Rating 4 :\",df[df['RATING']==3]['TEXT'].values[0])\r\n",
        "print(\"Rating 5 :\",df[df['RATING']==4]['TEXT'].values[0])\r\n",
        "print(\"------------------------------------------------------------------------------------------------------------------------------------------------------------------\")\r\n",
        "print(\"Rating 1 :\",df_clean[df_clean['RATING']==0]['TEXT'].values[0])\r\n",
        "print(\"Rating 2 :\",df_clean[df_clean['RATING']==1]['TEXT'].values[0])\r\n",
        "print(\"Rating 3 :\",df_clean[df_clean['RATING']==2]['TEXT'].values[0])\r\n",
        "print(\"Rating 4 :\",df_clean[df_clean['RATING']==3]['TEXT'].values[0])\r\n",
        "print(\"Rating 5 :\",df_clean[df_clean['RATING']==4]['TEXT'].values[0])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Rating 1 : The shirt was more of a smock. I expected a soft little cotton t shirt, but I suppose it\\'s my fault for not reading the description carefully. Also, the white shirt was completely see through. I returned mine.\"}},{r:{id:46830320,si:1,pi:10860142,mu:\"{f7dedd4c-02d2-49ba-9553-4f9be0bdb9e5}\n",
            "Rating 2 : I liked the shirt. it\\'s pretty. but I couldn\\'t fit my arms in the shirt. the sleeves are too tight. get a size bigger if you don\\'t have twig arms\"}},{r:{id:47933895,si:1,v:3,t:3,r:1,h:\"Not what I expected\n",
            "Rating 3 : I was really excited to see this in mail today, but got very disappointed right after trying it. I'm a petite 5 ft. 3\\\n",
            "Rating 4 : I really like this dress! It\\'s so comfortable, the colors are great for the summer, and it flows so nicely! I wore it to work &amp; to a night lounge dressed up with wedges! I got many compliments! \n",
            "Rating 5 : This dress is absolutely wonderful. Was afraid it might be too long for me, but the length is perfect! Dont even need to wear heels with it. Im 5\\'4 and i got it in S.\"}},{r:{id:74223533,si:1,pi:16960075,dp:true,mu:\"maharaibsyed013@gmail.com\n",
            "------------------------------------------------------------------------------------------------------------------------------------------------------------------\n",
            "Rating 1 : shirt smock expected soft little cotton shirt suppose it fault reading description carefully also white shirt completely see through returned minerid46830320si1pi10860142muf7dedd4c02d249ba95534f9be0bdb9e5\n",
            "Rating 2 : liked shirt it pretty couldnt fit arm shirt sleeve tight get size bigger dont twig armsrid47933895si1v3t3r1hnot expected\n",
            "Rating 3 : really excited see mail today got disappointed right trying it im petite 5 ft 3\n",
            "Rating 4 : really like dress it comfortable color great summer flow nicely wore work amp night lounge dressed wedge got many compliment\n",
            "Rating 5 : dress absolutely wonderful afraid might long me length perfect dont even need wear heel it im 54 got srid74223533si1pi16960075dptruemumaharaibsyed013gmailcom\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lW1L4V3i7hp_"
      },
      "source": [
        "Plot the predominant words of each class"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zUV2G3NX7p64"
      },
      "source": [
        "rating_1 = []\r\n",
        "for review in df_clean.loc[df_clean[\"RATING\"] == 0][\"TEXT\"]:\r\n",
        "  token = nltk.word_tokenize(review)\r\n",
        "  rating_1.extend(token)\r\n",
        "\r\n",
        "rating_2 = []\r\n",
        "for review in df_clean.loc[df_clean[\"RATING\"] == 1][\"TEXT\"]:\r\n",
        "  token = nltk.word_tokenize(review)\r\n",
        "  rating_2.extend(token)\r\n",
        "\r\n",
        "rating_3 = []\r\n",
        "for review in df_clean.loc[df_clean[\"RATING\"] == 2][\"TEXT\"]:\r\n",
        "  token = nltk.word_tokenize(review)\r\n",
        "  rating_3.extend(token)\r\n",
        "\r\n",
        "rating_4 = []\r\n",
        "for review in df_clean.loc[df_clean[\"RATING\"] == 3][\"TEXT\"]:\r\n",
        "  token = nltk.word_tokenize(review)\r\n",
        "  rating_4.extend(token)\r\n",
        "\r\n",
        "rating_5 = []\r\n",
        "for review in df_clean.loc[df_clean[\"RATING\"] == 4][\"TEXT\"]:\r\n",
        "  token = nltk.word_tokenize(review)\r\n",
        "  rating_5.extend(token)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "uOdM_FuB7okJ",
        "outputId": "2795fa92-a0b2-40a1-a41f-a2076a22d364"
      },
      "source": [
        "# Word frequency\r\n",
        "\r\n",
        "freq_1 = nltk.FreqDist(rating_1)\r\n",
        "freq_2 = nltk.FreqDist(rating_2)\r\n",
        "freq_3 = nltk.FreqDist(rating_3)\r\n",
        "freq_4 = nltk.FreqDist(rating_4)\r\n",
        "freq_5 = nltk.FreqDist(rating_5)\r\n",
        "\r\n",
        "freq_1.plot(5, cumulative=False, title=\"Rating 1/5\")\r\n",
        "freq_2.plot(5, cumulative=False, title=\"Rating 2/5\")\r\n",
        "freq_3.plot(5, cumulative=False, title=\"Rating 3/5\")\r\n",
        "freq_4.plot(5, cumulative=False, title=\"Rating 4/5\")\r\n",
        "freq_5.plot(5, cumulative=False, title=\"Rating 5/5\")\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IRLdHa1_Y-jg"
      },
      "source": [
        "### Training our model using Transfer Learning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QgMvyEm8edAZ"
      },
      "source": [
        "Import BERT from the transformers library"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 232,
          "referenced_widgets": [
            "93dd867f6c424abf986a39e6d16aa389",
            "92b219d31d89435da2793956431c0215",
            "e6df8030d3ca4e85a4d197d7c1268d78",
            "2c7c06e8e7374c33a3433765dad28667",
            "9191dad4453544388092bdac04522114",
            "0b3ca3afad23462180cc47208227d0f9",
            "d3da291fcd784edea6b3db29c799cfa0",
            "4ce78e8e3eea45b5b2b815515f64afd7",
            "de2f20f766bd40b39f8c3cce077d1207",
            "528e15a242e140a7aea4377f3a8860e9",
            "d146333b63624bfe98163ad89bdcaaf3",
            "0f340f2eca6749188812e2f212b94b1b",
            "2f3ada5798734d73b3b99306c14d28d0",
            "6eb0d451d694459194022a982c504dc3",
            "6478c92973bb4e0d933af010e865ef59",
            "b1dd7340424c4611aeefc485d5ee5358",
            "9320ce02c1684ef29557b7ce9bae9e7d",
            "94d6964281914b029c70996a727c5092",
            "08f4f7a1d20f420eaf8289c3ca4ddcac",
            "eeb1247c1db2481daff83dc07d98a949",
            "a1148332a53e4561ad7048d549a77599",
            "8a48d78aa20540998b3673b6255d55da",
            "c19805df136b40c286903770668c3742",
            "dae8d41fb7d2486c87083726a34a8ede"
          ]
        },
        "id": "bueOt3Y_egb3",
        "outputId": "cbe29ade-044e-49d1-cb36-ebe80a825c7a"
      },
      "source": [
        "from transformers import BertConfig, BertTokenizer, TFBertForSequenceClassification, InputExample, InputFeatures\r\n",
        "\r\n",
        "config = BertConfig.from_pretrained(\"bert-base-uncased\", hidden_dropout_prob=0.5, num_labels=5)\r\n",
        "model = TFBertForSequenceClassification.from_pretrained(\"bert-base-uncased\", config=config)\r\n",
        "tokenizer = BertTokenizer.from_pretrained(\"bert-base-uncased\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "93dd867f6c424abf986a39e6d16aa389",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=433.0, style=ProgressStyle(description_…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "de2f20f766bd40b39f8c3cce077d1207",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=536063208.0, style=ProgressStyle(descri…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "All model checkpoint layers were used when initializing TFBertForSequenceClassification.\n",
            "\n",
            "Some layers of TFBertForSequenceClassification were not initialized from the model checkpoint at bert-base-uncased and are newly initialized: ['classifier']\n",
            "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "9320ce02c1684ef29557b7ce9bae9e7d",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=231508.0, style=ProgressStyle(descripti…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JOGMQdlfeoLQ",
        "outputId": "91feb649-8042-4b21-9d40-ce45c998265e"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"tf_bert_for_sequence_classification\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "bert (TFBertMainLayer)       multiple                  109482240 \n",
            "_________________________________________________________________\n",
            "dropout_37 (Dropout)         multiple                  0         \n",
            "_________________________________________________________________\n",
            "classifier (Dense)           multiple                  3845      \n",
            "=================================================================\n",
            "Total params: 109,486,085\n",
            "Trainable params: 109,486,085\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KXHI2TEAgPlf"
      },
      "source": [
        "Train Test Split"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DPlDPt35gR5l"
      },
      "source": [
        "from sklearn.model_selection import train_test_split\r\n",
        "\r\n",
        "train, test = train_test_split(df_clean, test_size=0.25, random_state=123, shuffle=True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fbSGHQU7uQbD",
        "outputId": "c0d78fb0-0a51-47e5-f9ac-3081b64ecd27"
      },
      "source": [
        "train.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2600, 2)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Q6q4vnVluQZI",
        "outputId": "91c208fe-2600-48c1-f1f5-43f0b023ff3d"
      },
      "source": [
        "test.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(867, 2)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nh-jQXyRkS4R"
      },
      "source": [
        "I'll use BERT's InputExample method to help us create sequences from our dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jYe8SULRgR3T",
        "outputId": "dd0e3565-b2c2-4207-a2c4-96bcd9a85e10"
      },
      "source": [
        "InputExample(guid = None,\r\n",
        "             text_a = \"Hello, world\",\r\n",
        "             text_b = None,\r\n",
        "             label = 1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "InputExample(guid=None, text_a='Hello, world', text_b=None, label=1)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Jod7xhsZM_95"
      },
      "source": [
        "def convert_data_to_examples(train, test, DATA_COLUMN, LABEL_COLUMN): \r\n",
        "  train_InputExamples = train.apply(lambda x: InputExample(guid=None, # Globally unique ID for bookkeeping, unused in this case\r\n",
        "                                                          text_a = x[DATA_COLUMN], \r\n",
        "                                                          text_b = None,\r\n",
        "                                                          label = x[LABEL_COLUMN]), axis = 1)\r\n",
        "\r\n",
        "  validation_InputExamples = test.apply(lambda x: InputExample(guid=None, # Globally unique ID for bookkeeping, unused in this case\r\n",
        "                                                          text_a = x[DATA_COLUMN], \r\n",
        "                                                          text_b = None,\r\n",
        "                                                          label = x[LABEL_COLUMN]), axis = 1)\r\n",
        "  \r\n",
        "  return train_InputExamples, validation_InputExamples"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bvhI-51TgR0u"
      },
      "source": [
        "def convert_examples_to_tf_dataset(examples, tokenizer, max_length=128):\r\n",
        "    features = [] # -> will hold InputFeatures to be converted later\r\n",
        "\r\n",
        "    for e in examples:\r\n",
        "        # Documentation is really strong for this method, so please take a look at it\r\n",
        "        input_dict = tokenizer.encode_plus(\r\n",
        "            e.text_a,\r\n",
        "            add_special_tokens=True,\r\n",
        "            max_length=max_length, # truncates if len(s) > max_length\r\n",
        "            return_token_type_ids=True,\r\n",
        "            return_attention_mask=True,\r\n",
        "            pad_to_max_length=True, # pads to the right by default # CHECK THIS for pad_to_max_length\r\n",
        "            truncation=True\r\n",
        "        )\r\n",
        "\r\n",
        "        input_ids, token_type_ids, attention_mask = (input_dict[\"input_ids\"],\r\n",
        "            input_dict[\"token_type_ids\"], input_dict['attention_mask'])\r\n",
        "\r\n",
        "        features.append(\r\n",
        "            InputFeatures(\r\n",
        "                input_ids=input_ids, attention_mask=attention_mask, token_type_ids=token_type_ids, label=e.label\r\n",
        "            )\r\n",
        "        )\r\n",
        "\r\n",
        "    def gen():\r\n",
        "        for f in features:\r\n",
        "            yield (\r\n",
        "                {\r\n",
        "                    \"input_ids\": f.input_ids,\r\n",
        "                    \"attention_mask\": f.attention_mask,\r\n",
        "                    \"token_type_ids\": f.token_type_ids,\r\n",
        "                },\r\n",
        "                f.label,\r\n",
        "            )\r\n",
        "\r\n",
        "    return tf.data.Dataset.from_generator(gen,\r\n",
        "                                          ({\"input_ids\": tf.int32, \"attention_mask\": tf.int32, \"token_type_ids\": tf.int32}, tf.int64),\r\n",
        "                                          (\r\n",
        "                                            {\r\n",
        "                                                \"input_ids\": tf.TensorShape([None]),\r\n",
        "                                                \"attention_mask\": tf.TensorShape([None]),\r\n",
        "                                                \"token_type_ids\": tf.TensorShape([None]),\r\n",
        "                                            },\r\n",
        "                                            tf.TensorShape([]),\r\n",
        "                                          ),\r\n",
        "                                         )"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aKbJhZf8NmLO"
      },
      "source": [
        "DATA_COLUMN = 'TEXT'\r\n",
        "LABEL_COLUMN = 'RATING'"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dTu1O9uFm6bT",
        "outputId": "c1e66f1e-6093-4f02-c587-853f5781444f"
      },
      "source": [
        "train_InputExamples, validation_InputExamples = convert_data_to_examples(train, test, DATA_COLUMN, LABEL_COLUMN)\r\n",
        "\r\n",
        "train_data = convert_examples_to_tf_dataset(list(train_InputExamples), tokenizer)\r\n",
        "train_data = train_data.shuffle(100).batch(32).repeat(2)\r\n",
        "\r\n",
        "validation_data = convert_examples_to_tf_dataset(list(validation_InputExamples), tokenizer)\r\n",
        "validation_data = validation_data.batch(32)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/transformers/tokenization_utils_base.py:2155: FutureWarning: The `pad_to_max_length` argument is deprecated and will be removed in a future version, use `padding=True` or `padding='longest'` to pad to the longest sequence in the batch, or use `padding='max_length'` to pad to a max length. In this case, you can give a specific length with `max_length` (e.g. `max_length=45`) or leave max_length to None to pad to the maximal input size of the model (e.g. 512 for Bert).\n",
            "  FutureWarning,\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IWqLypYX7ubq",
        "outputId": "e5fd74f5-732e-4738-b06a-4b7a625e33b8"
      },
      "source": [
        "list(train_InputExamples[:5]) #to see the structure of our data using the InputExample BERT function"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[InputExample(guid=None, text_a='bustier make double pumping much easier great idea zipper one broke within first two week working keep going would recommend going another brand worth money hassle', text_b=None, label=1),\n",
              " InputExample(guid=None, text_a='fabric dress substantial feel high quality construction done well however fit dress odd short im 55 135 lb im small target dress thats ordered dress suppose correct size fit everywhere however short unless shorter am would absolutely consider work dress waist hit much higher natural waist again petite height might work you agree previous reviewer way skirt flare flattering add lot bulk high hope dress bomb', text_b=None, label=1),\n",
              " InputExample(guid=None, text_a='quite surprised styling versatility modal lounge pant 59 wa 335 in inseam 235lbs size 1820 somewhat curvy ordered long xxl im glad length perfect pant quite long though usually roll waistline tie bring legnth wearing flat walking shoesgym shoe tie regular waistline wearing high heeled booty dress boot super versatile styling well wear casual pant dressy pant one know loungewearnnthese pant thin show bump bulge easily material substantially roomy pant leg wide flareleg style would definitely wear spring summer weather u really want wear fall winter month put legging underneath keep warmnnbased legnth long pant see anyone 36 in inseam rockin lounge pant problem bought sale buy one get 60 another pair use 20 promo ordermade pant steal black brown pair look nice dressy gray pair casual heather gray look for walkingworking out navy blue dark navy blue wash quite wellive already washed warm hot water thrown dryer high heat shrunk yetnngreat buy especially get sale really comfortable versatile lounge pant gave 4stars though pretty thin light good buy none le', text_b=None, label=3),\n",
              " InputExample(guid=None, text_a='strap short even way out elastic used scratchy tried wearing today washing 2 time could hardly wait get home take off', text_b=None, label=0),\n",
              " InputExample(guid=None, text_a='generally fan underwires really needed minimizer bra wear lighter summer clothing bra surprisingly comfy great job minimizing also true size', text_b=None, label=3)]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dTnHaxYF8L9p",
        "outputId": "94bdea7c-b5d9-486a-c2cf-436240201cea"
      },
      "source": [
        "list(train_data.take(1)) #to see the final structure of our dataset, generated by TensorFlow's tf.data.Dataset.from_generator() method"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[({'attention_mask': <tf.Tensor: shape=(32, 128), dtype=int32, numpy=\n",
              "   array([[1, 1, 1, ..., 0, 0, 0],\n",
              "          [1, 1, 1, ..., 0, 0, 0],\n",
              "          [1, 1, 1, ..., 0, 0, 0],\n",
              "          ...,\n",
              "          [1, 1, 1, ..., 0, 0, 0],\n",
              "          [1, 1, 1, ..., 0, 0, 0],\n",
              "          [1, 1, 1, ..., 0, 0, 0]], dtype=int32)>,\n",
              "   'input_ids': <tf.Tensor: shape=(32, 128), dtype=int32, numpy=\n",
              "   array([[  101,  4906, 17099, ...,     0,     0,     0],\n",
              "          [  101,  4156,  7579, ...,     0,     0,     0],\n",
              "          [  101,  3191,  3319, ...,     0,     0,     0],\n",
              "          ...,\n",
              "          [  101,  4606,  2946, ...,     0,     0,     0],\n",
              "          [  101,  2364, 10520, ...,     0,     0,     0],\n",
              "          [  101,  3517,  2438, ...,     0,     0,     0]], dtype=int32)>,\n",
              "   'token_type_ids': <tf.Tensor: shape=(32, 128), dtype=int32, numpy=\n",
              "   array([[0, 0, 0, ..., 0, 0, 0],\n",
              "          [0, 0, 0, ..., 0, 0, 0],\n",
              "          [0, 0, 0, ..., 0, 0, 0],\n",
              "          ...,\n",
              "          [0, 0, 0, ..., 0, 0, 0],\n",
              "          [0, 0, 0, ..., 0, 0, 0],\n",
              "          [0, 0, 0, ..., 0, 0, 0]], dtype=int32)>},\n",
              "  <tf.Tensor: shape=(32,), dtype=int64, numpy=\n",
              "  array([3, 4, 0, 1, 3, 2, 3, 1, 2, 3, 1, 4, 1, 2, 1, 2, 1, 3, 0, 3, 3, 3,\n",
              "         4, 3, 3, 2, 4, 2, 3, 0, 1, 2])>)]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 35
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1fI8xmrwm9OO"
      },
      "source": [
        "Compile and train the model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "V0s86DQ1QWGS",
        "outputId": "6810c332-6358-4bb2-dc88-e5fe6cf1214d"
      },
      "source": [
        "import os\r\n",
        "from keras.callbacks import EarlyStopping, ModelCheckpoint\r\n",
        "#checkpoint_filepath = os.path.join(os.getcwd(), \"checkpoints/weights-cnn.hdf5\")\r\n",
        "from google.colab import drive\r\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vRGUtJOpgRya",
        "outputId": "fe97456c-5892-48e7-e685-a5ab67c1b5e1"
      },
      "source": [
        "model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=3e-5, epsilon=1e-08, clipnorm=1.0), \r\n",
        "              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True), \r\n",
        "              metrics=[tf.keras.metrics.SparseCategoricalAccuracy('accuracy')])\r\n",
        "\r\n",
        "checkpoint_filepath = '/content/drive/My Drive/Colab Notebooks/weights-nlp.hdf5'\r\n",
        "checkpoint = ModelCheckpoint(checkpoint_filepath, monitor='val_loss', verbose=1, save_best_only=True, mode='min')\r\n",
        "\r\n",
        "history = model.fit(train_data, epochs=5, validation_data=validation_data, callbacks=checkpoint)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "WARNING:tensorflow:The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model.They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`).\n",
            "WARNING:tensorflow:AutoGraph could not transform <bound method Socket.send of <zmq.sugar.socket.Socket object at 0x7f7d87642b40>> and will run it as-is.\n",
            "Please report this to the TensorFlow team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output.\n",
            "Cause: module, class, method, function, traceback, frame, or code object was expected, got cython_function_or_method\n",
            "To silence this warning, decorate the function with @tf.autograph.experimental.do_not_convert\n",
            "WARNING: AutoGraph could not transform <bound method Socket.send of <zmq.sugar.socket.Socket object at 0x7f7d87642b40>> and will run it as-is.\n",
            "Please report this to the TensorFlow team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output.\n",
            "Cause: module, class, method, function, traceback, frame, or code object was expected, got cython_function_or_method\n",
            "To silence this warning, decorate the function with @tf.autograph.experimental.do_not_convert\n",
            "WARNING:tensorflow:AutoGraph could not transform <function wrap at 0x7f7da29ea0e0> and will run it as-is.\n",
            "Cause: while/else statement not yet supported\n",
            "To silence this warning, decorate the function with @tf.autograph.experimental.do_not_convert\n",
            "WARNING: AutoGraph could not transform <function wrap at 0x7f7da29ea0e0> and will run it as-is.\n",
            "Cause: while/else statement not yet supported\n",
            "To silence this warning, decorate the function with @tf.autograph.experimental.do_not_convert\n",
            "WARNING:tensorflow:The parameter `return_dict` cannot be set in graph mode and will always be set to `True`.\n",
            "WARNING:tensorflow:The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model.They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`).\n",
            "WARNING:tensorflow:The parameter `return_dict` cannot be set in graph mode and will always be set to `True`.\n",
            "    164/Unknown - 7708s 47s/step - loss: 1.6389 - accuracy: 0.2201WARNING:tensorflow:The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model.They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`).\n",
            "WARNING:tensorflow:The parameter `return_dict` cannot be set in graph mode and will always be set to `True`.\n",
            "164/164 [==============================] - 8099s 49s/step - loss: 1.6388 - accuracy: 0.2201 - val_loss: 1.6084 - val_accuracy: 0.2376\n",
            "\n",
            "Epoch 00001: val_loss improved from inf to 1.60842, saving model to /content/drive/My Drive/Colab Notebooks/weights-nlp.hdf5\n",
            "Epoch 2/5\n",
            "164/164 [==============================] - 8140s 50s/step - loss: 1.6129 - accuracy: 0.2294 - val_loss: 1.6078 - val_accuracy: 0.2238\n",
            "\n",
            "Epoch 00002: val_loss improved from 1.60842 to 1.60778, saving model to /content/drive/My Drive/Colab Notebooks/weights-nlp.hdf5\n",
            "Epoch 3/5\n",
            "111/164 [===================>..........] - ETA: 42:06 - loss: 1.6082 - accuracy: 0.2403"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_eDxY0sbZloP"
      },
      "source": [
        "### Final Thoughts:"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IEnTJm33i1KQ"
      },
      "source": [
        "* I've cleaned the data making text lowercase, removing punctuation, numbers and stopwords, and performing lemmatization. I've used nltk library to remove stopwords and the lemmatization.\r\n",
        "\r\n",
        "* For the learning process I've used the clean lemmatized text. I've chosen the BERT model from the transformers library, as it's the state of art model for NLP and Transfer Learning. I've also used TensorFlow, for the input of the model.\r\n",
        "\r\n",
        "* For the model evaluation, I've used the accuracy metric and the SparseCategoricalCrossentropy for the loss function. For the validation set, I've split the dataset at 75%/25%. The testing accuracy is 24%. It's too low, but I wanted to use the BERT model which has barely been published as open source for these type of sentiment analysis, which classifies not as positive/neutral/negative, but on 5 rating classes. It also needs a lot of hours for training. \r\n",
        "\r\n",
        "* With more time and data, I would have tried another models or training this one with more epochs, as well as training a RNN using Keras. Furthermore, 3400 samples are too few for a neural network.\r\n",
        "\r\n",
        "* Future work: change the ratings from 1-5 to negative (1-2), neutral (3) and positive (4-5) to simplify the training."
      ]
    }
  ]
}