Update docs

_downloads/2421e198bfdefccbcbfa5427d1acaa31/plot_example1.ipynb

@@ -15,7 +15,7 @@
       "cell_type": "markdown",
       "metadata": {},
       "source": [
-        "\n\n# Example: Use of MRC with different settings\n\nExample of using MRC with some of the common classification datasets with different\nlosses and feature mappings settings. We load the different datasets and use 10-Fold \nCross-Validation to generate the partitions for train and test. We separate 1 partition\neach time for testing and use the others for training. On each iteration we calculate\nthe classification error as well as the upper and lower bounds for the error. We also\ncalculate the mean training time.\n\nYou can check a more elaborated example in `ex_comp`.\n"
+        "\n\n# Example: Use of MRC with different settings\n\nExample of using MRC with some of the common classification datasets with\ndifferent losses and feature mappings settings. We load the different datasets\nand use 10-Fold Cross-Validation to generate the partitions for train and test.\nWe separate 1 partition each time for testing and use the others for training.\nOn each iteration we calculate the classification error as well as the upper\nand lower bounds for the error. We also\ncalculate the mean training time.\n\nYou can check a more elaborated example in `ex_comp`.\n"
       ]
     },
     {

_downloads/a8b175cd29751dc0229ceb76947711bf/plot_example2.py

@@ -5,10 +5,11 @@
 Example: Use of CMRC with different settings
 ============================================
 
-Example of using CMRC with some of the common classification datasets with different
-losses and feature mappings settings. We load the different datasets and use 10-Fold 
-Cross-Validation to generate the partitions for train and test. We separate 1 partition
-each time for testing and use the others for training. On each iteration we calculate 
+Example of using CMRC with some of the common classification datasets with
+different losses and feature mappings settings. We load the different datasets
+and use 10-Fold Cross-Validation to generate the partitions for train and test.
+We separate 1 partition each time for testing and use the others for training.
+On each iteration we calculate
 the classification error. We also calculate the mean training time.
 
 You can check a more elaborated example in :ref:`ex_comp`.
@@ -27,7 +28,7 @@
 # Data sets
 loaders = [load_mammographic, load_haberman, load_indian_liver,
            load_diabetes, load_credit]
-dataName = ["mammographic", "haberman", "indian_liver", 
+dataName = ["mammographic", "haberman", "indian_liver",
             "diabetes", "credit"]
 
 
@@ -82,7 +83,7 @@ def runCMRC(phi, loss):
 
             # Save the training time
             auxTime += time.time() - startTime
-            
+
             # Predict the class for test instances
             y_pred = clf.predict(X_test)
 
@@ -102,7 +103,8 @@ def runCMRC(phi, loss):
 
 if __name__ == '__main__':
 
-    print('*** Example (CMRC with the additional marginal constraints) *** \n\n')
+    print('*** Example (CMRC with the additional\
+     marginal constraints) *** \n\n')
 
     print('1. Using 0-1 loss and relu feature mapping \n\n')
     runCMRC(phi='relu', loss='0-1')

_downloads/cd55f92c8d819b776c6a1e904e5efb23/plot_example2.ipynb

@@ -15,7 +15,7 @@
       "cell_type": "markdown",
       "metadata": {},
       "source": [
-        "\n\n# Example: Use of CMRC with different settings\n\nExample of using CMRC with some of the common classification datasets with different\nlosses and feature mappings settings. We load the different datasets and use 10-Fold \nCross-Validation to generate the partitions for train and test. We separate 1 partition\neach time for testing and use the others for training. On each iteration we calculate \nthe classification error. We also calculate the mean training time.\n\nYou can check a more elaborated example in `ex_comp`.\n"
+        "\n\n# Example: Use of CMRC with different settings\n\nExample of using CMRC with some of the common classification datasets with\ndifferent losses and feature mappings settings. We load the different datasets\nand use 10-Fold Cross-Validation to generate the partitions for train and test.\nWe separate 1 partition each time for testing and use the others for training.\nOn each iteration we calculate\nthe classification error. We also calculate the mean training time.\n\nYou can check a more elaborated example in `ex_comp`.\n"
       ]
     },
     {
@@ -26,7 +26,7 @@
       },
       "outputs": [],
       "source": [
-        "import time\n\nimport numpy as np\nfrom sklearn import preprocessing\nfrom sklearn.model_selection import StratifiedKFold\n\nfrom MRCpy import CMRC\n# Import the datasets\nfrom MRCpy.datasets import *\n\n# Data sets\nloaders = [load_mammographic, load_haberman, load_indian_liver,\n           load_diabetes, load_credit]\ndataName = [\"mammographic\", \"haberman\", \"indian_liver\", \n            \"diabetes\", \"credit\"]\n\n\ndef runCMRC(phi, loss):\n\n    res_mean = np.zeros(len(dataName))\n    res_std = np.zeros(len(dataName))\n\n    # We fix the random seed to that the stratified kfold performed\n    # is the same through the different executions\n    random_seed = 0\n\n    # Iterate through each of the dataset and fit the CMRC classfier.\n    for j, load in enumerate(loaders):\n\n        # Loading the dataset\n        X, Y = load(return_X_y=True)\n        r = len(np.unique(Y))\n        n, d = X.shape\n\n        # Print the dataset name\n        print(\" ############## \\n \" + dataName[j] + \" n= \" + str(n) +\n              \" , d= \" + str(d) + \", cardY= \" + str(r))\n\n        # Create the CMRC object initilized with the corresponding parameters\n        clf = CMRC(phi=phi, loss=loss, use_cvx=True,\n                   solver='MOSEK', max_iters=10000, s=0.3)\n\n        # Generate the partitions of the stratified cross-validation\n        cv = StratifiedKFold(n_splits=10, random_state=random_seed,\n                             shuffle=True)\n\n        cvError = list()\n        auxTime = 0\n\n        # Paired and stratified cross-validation\n        for train_index, test_index in cv.split(X, Y):\n\n            X_train, X_test = X[train_index], X[test_index]\n            y_train, y_test = Y[train_index], Y[test_index]\n\n            # Normalizing the data\n            std_scale = preprocessing.StandardScaler().fit(X_train, y_train)\n            X_train = std_scale.transform(X_train)\n            X_test = std_scale.transform(X_test)\n\n            # Save start time for computing training time\n            startTime = time.time()\n\n            # Train the model\n            clf.fit(X_train, y_train)\n\n            # Save the training time\n            auxTime += time.time() - startTime\n            \n            # Predict the class for test instances\n            y_pred = clf.predict(X_test)\n\n            # Calculate the error made by CMRC classificator\n            cvError.append(np.average(y_pred != y_test))\n\n        res_mean[j] = np.average(cvError)\n        res_std[j] = np.std(cvError)\n\n        # Calculating the mean training time\n        auxTime = auxTime / 10\n\n        print(\" error= \" + \": \" + str(res_mean[j]) + \" +/- \" +\n              str(res_std[j]) + \"\\n avg_train_time= \" + \": \" +\n              str(auxTime) + ' secs' + \"\\n ############## \\n\\n\")\n\n\nif __name__ == '__main__':\n\n    print('*** Example (CMRC with the additional marginal constraints) *** \\n\\n')\n\n    print('1. Using 0-1 loss and relu feature mapping \\n\\n')\n    runCMRC(phi='relu', loss='0-1')\n\n    print('2. Using log loss and relu feature mapping \\n\\n')\n    runCMRC(phi='relu', loss='log')"
+        "import time\n\nimport numpy as np\nfrom sklearn import preprocessing\nfrom sklearn.model_selection import StratifiedKFold\n\nfrom MRCpy import CMRC\n# Import the datasets\nfrom MRCpy.datasets import *\n\n# Data sets\nloaders = [load_mammographic, load_haberman, load_indian_liver,\n           load_diabetes, load_credit]\ndataName = [\"mammographic\", \"haberman\", \"indian_liver\",\n            \"diabetes\", \"credit\"]\n\n\ndef runCMRC(phi, loss):\n\n    res_mean = np.zeros(len(dataName))\n    res_std = np.zeros(len(dataName))\n\n    # We fix the random seed to that the stratified kfold performed\n    # is the same through the different executions\n    random_seed = 0\n\n    # Iterate through each of the dataset and fit the CMRC classfier.\n    for j, load in enumerate(loaders):\n\n        # Loading the dataset\n        X, Y = load(return_X_y=True)\n        r = len(np.unique(Y))\n        n, d = X.shape\n\n        # Print the dataset name\n        print(\" ############## \\n \" + dataName[j] + \" n= \" + str(n) +\n              \" , d= \" + str(d) + \", cardY= \" + str(r))\n\n        # Create the CMRC object initilized with the corresponding parameters\n        clf = CMRC(phi=phi, loss=loss, use_cvx=True,\n                   solver='MOSEK', max_iters=10000, s=0.3)\n\n        # Generate the partitions of the stratified cross-validation\n        cv = StratifiedKFold(n_splits=10, random_state=random_seed,\n                             shuffle=True)\n\n        cvError = list()\n        auxTime = 0\n\n        # Paired and stratified cross-validation\n        for train_index, test_index in cv.split(X, Y):\n\n            X_train, X_test = X[train_index], X[test_index]\n            y_train, y_test = Y[train_index], Y[test_index]\n\n            # Normalizing the data\n            std_scale = preprocessing.StandardScaler().fit(X_train, y_train)\n            X_train = std_scale.transform(X_train)\n            X_test = std_scale.transform(X_test)\n\n            # Save start time for computing training time\n            startTime = time.time()\n\n            # Train the model\n            clf.fit(X_train, y_train)\n\n            # Save the training time\n            auxTime += time.time() - startTime\n\n            # Predict the class for test instances\n            y_pred = clf.predict(X_test)\n\n            # Calculate the error made by CMRC classificator\n            cvError.append(np.average(y_pred != y_test))\n\n        res_mean[j] = np.average(cvError)\n        res_std[j] = np.std(cvError)\n\n        # Calculating the mean training time\n        auxTime = auxTime / 10\n\n        print(\" error= \" + \": \" + str(res_mean[j]) + \" +/- \" +\n              str(res_std[j]) + \"\\n avg_train_time= \" + \": \" +\n              str(auxTime) + ' secs' + \"\\n ############## \\n\\n\")\n\n\nif __name__ == '__main__':\n\n    print('*** Example (CMRC with the additional\\\n     marginal constraints) *** \\n\\n')\n\n    print('1. Using 0-1 loss and relu feature mapping \\n\\n')\n    runCMRC(phi='relu', loss='0-1')\n\n    print('2. Using log loss and relu feature mapping \\n\\n')\n    runCMRC(phi='relu', loss='log')"
       ]
     }
   ],

_downloads/ce615eeeaf183d728a05b647cb915a04/plot_example1.py

@@ -5,11 +5,12 @@
 Example: Use of MRC with different settings
 ===========
 
-Example of using MRC with some of the common classification datasets with different
-losses and feature mappings settings. We load the different datasets and use 10-Fold 
-Cross-Validation to generate the partitions for train and test. We separate 1 partition
-each time for testing and use the others for training. On each iteration we calculate
-the classification error as well as the upper and lower bounds for the error. We also
+Example of using MRC with some of the common classification datasets with
+different losses and feature mappings settings. We load the different datasets
+and use 10-Fold Cross-Validation to generate the partitions for train and test.
+We separate 1 partition each time for testing and use the others for training.
+On each iteration we calculate the classification error as well as the upper
+and lower bounds for the error. We also
 calculate the mean training time.
 
 You can check a more elaborated example in :ref:`ex_comp`.