{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# Introduction to Target Encoding","metadata":{}},{"cell_type":"markdown","source":"This is an introduction to Target Encoding, exhibiting the application of K-Fold Target Encoding technique on the Titanic competition data.","metadata":{}},{"cell_type":"markdown","source":"![image.png](attachment:image.png)","metadata":{},"attachments":{"image.png":{"image/png":"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"}}},{"cell_type":"markdown","source":"Target Encoding is similar to label encoding, except here labels are correlated directly with the target. Mean target encoding for each category in the feature label is decided with the mean value of the target variable on a training data. This encoding method brings out the relation between similar categories, but the relations are bounded within the categories and target itself. \n\nThe advantages of the mean target encoding are that it does not affect the volume of the data and helps in faster learning. Many competition winning models make use of target encoding as it often produces higher accuracy (sometimes drastically high accuracies) because of the direct correlation between the encoded variable and the target. \n\nOverfitting can be a major problem while using target encoding which can be addressed by using the K-Fold Target Encoding technique.","metadata":{}},{"cell_type":"markdown","source":"![image.png](attachment:image.png)","metadata":{},"attachments":{"image.png":{"image/png":"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"}}},{"cell_type":"markdown","source":"The idea is similar to k-fold cross validation. We divide the data in K-stratified or random folds, replace the observations present in the M-th fold with mean target of data from all other folds except the M-th fold. We are basically trying to use all the data given to us and not leak the information from self target label by allowing target information to flow from other fellow observations (same category but other folds).","metadata":{}},{"cell_type":"markdown","source":"# The Code for application","metadata":{}},{"cell_type":"markdown","source":"We will now apply K-Fold Target Enoding technique on the Titanic competition data by encoding the **Embarked feature** of the dataset.","metadata":{}},{"cell_type":"code","source":"import pandas as pd\nimport numpy as np\nimport matplotlib.pyplot as plt\nfrom sklearn import base\nfrom sklearn.model_selection import KFold\n\ntrain = pd.read_csv(\"../input/titanic/train.csv\")\ntest = pd.read_csv(\"../input/titanic/test.csv\")","metadata":{"execution":{"iopub.status.busy":"2022-09-19T11:10:08.500935Z","iopub.execute_input":"2022-09-19T11:10:08.501298Z","iopub.status.idle":"2022-09-19T11:10:09.544381Z","shell.execute_reply.started":"2022-09-19T11:10:08.501266Z","shell.execute_reply":"2022-09-19T11:10:09.543268Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"First, a k-fold target encoder class will be created for the train set. The KFoldTargetEncoderTrain class gets the name of the feature column, target column and number of fold initially and then fit and transform the train data. ","metadata":{}},{"cell_type":"code","source":"class KFoldTargetEncoderTrain(base.BaseEstimator,\n                               base.TransformerMixin):\n    def __init__(self,colnames,targetName,\n                  n_fold=5, verbosity=True,\n                  discardOriginal_col=False):\n        self.colnames = colnames\n        self.targetName = targetName\n        self.n_fold = n_fold\n        self.verbosity = verbosity\n        self.discardOriginal_col = discardOriginal_col\n    def fit(self, X, y=None):\n        return self\n    def transform(self,X):\n        mean_of_target = X[self.targetName].mean()\n        kf = KFold(n_splits = self.n_fold,\n                   shuffle = False, random_state=2019)\n        col_mean_name = self.colnames + '_' + 'Kfold_Target_Enc'\n        X[col_mean_name] = np.nan\n        for tr_ind, val_ind in kf.split(X):\n            X_tr, X_val = X.iloc[tr_ind], X.iloc[val_ind]\n            X.loc[X.index[val_ind], col_mean_name] = X_val[self.colnames].map(X_tr.groupby(self.colnames)\n                                     [self.targetName].mean())\n            X[col_mean_name].fillna(mean_of_target, inplace = True)\n        if self.verbosity:\n            encoded_feature = X[col_mean_name].values\n            print('Correlation between the new feature, {} and, {} is {}.'.format(col_mean_name,self.targetName,                    \n                   np.corrcoef(X[self.targetName].values,\n                               encoded_feature)[0][1]))\n        if self.discardOriginal_col:\n            X = X.drop(self.targetName, axis=1)\n        return X","metadata":{"execution":{"iopub.status.busy":"2022-09-19T11:10:09.546438Z","iopub.execute_input":"2022-09-19T11:10:09.546791Z","iopub.status.idle":"2022-09-19T11:10:09.560955Z","shell.execute_reply.started":"2022-09-19T11:10:09.546756Z","shell.execute_reply":"2022-09-19T11:10:09.559931Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Now, we will create an object of this class and use it to fit and transform the train data. As mentioned earlier, we will encode the Embarked column of the dataset.","metadata":{}},{"cell_type":"code","source":"targetc = KFoldTargetEncoderTrain('Embarked','Survived',n_fold=5)\nnew_train = targetc.fit_transform(train)","metadata":{"execution":{"iopub.status.busy":"2022-09-19T11:10:09.562471Z","iopub.execute_input":"2022-09-19T11:10:09.562865Z","iopub.status.idle":"2022-09-19T11:10:09.661931Z","shell.execute_reply.started":"2022-09-19T11:10:09.562830Z","shell.execute_reply":"2022-09-19T11:10:09.661135Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Taking a look at the new_train dataset, we can see a new column named Embarked_Kfold_Target_Enc which is the target encoded version of the Embarked feature.","metadata":{}},{"cell_type":"code","source":"new_train.head()","metadata":{"execution":{"iopub.status.busy":"2022-09-19T11:10:09.663155Z","iopub.execute_input":"2022-09-19T11:10:09.663653Z","iopub.status.idle":"2022-09-19T11:10:09.692261Z","shell.execute_reply.started":"2022-09-19T11:10:09.663609Z","shell.execute_reply":"2022-09-19T11:10:09.691369Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Finally, we need to create the “Embarked_Kfold_Target_Enc” in the test dataset by using the following class. The class needs the train dataset and “Feature” (\"Embarked\") and the name of the encoded column. After that we fit and transform the test dataset.","metadata":{}},{"cell_type":"code","source":"class KFoldTargetEncoderTest(base.BaseEstimator, base.TransformerMixin):\n    \n    def __init__(self,train,colNames,encodedName):\n        \n        self.train = train\n        self.colNames = colNames\n        self.encodedName = encodedName\n        \n    def fit(self, X, y=None):\n        return self\n    def transform(self,X):\n        mean =  self.train[[self.colNames,\n                self.encodedName]].groupby(\n                                self.colNames).mean().reset_index() \n        \n        dd = {}\n        for index, row in mean.iterrows():\n            dd[row[self.colNames]] = row[self.encodedName]\n        X[self.encodedName] = X[self.colNames]\n        X = X.replace({self.encodedName: dd})\n        return X","metadata":{"execution":{"iopub.status.busy":"2022-09-19T11:10:09.694618Z","iopub.execute_input":"2022-09-19T11:10:09.694962Z","iopub.status.idle":"2022-09-19T11:10:09.704513Z","shell.execute_reply.started":"2022-09-19T11:10:09.694931Z","shell.execute_reply":"2022-09-19T11:10:09.703393Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Now, we will create an object of this class and use it to fit and transform the test data.","metadata":{}},{"cell_type":"code","source":"test_targetc = KFoldTargetEncoderTest(new_train,\n                                      'Embarked',\n                                      'Embarked_Kfold_Target_Enc')\nnew_test = test_targetc.fit_transform(test)","metadata":{"execution":{"iopub.status.busy":"2022-09-19T11:10:09.706467Z","iopub.execute_input":"2022-09-19T11:10:09.707184Z","iopub.status.idle":"2022-09-19T11:10:09.732158Z","shell.execute_reply.started":"2022-09-19T11:10:09.707139Z","shell.execute_reply":"2022-09-19T11:10:09.731210Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Taking a look at the new_test dataset, we can see a new column named Embarked_Kfold_Target_Enc which is the target encoded version of the Embarked feature.","metadata":{}},{"cell_type":"code","source":"new_test.head()","metadata":{"execution":{"iopub.status.busy":"2022-09-19T11:10:09.733754Z","iopub.execute_input":"2022-09-19T11:10:09.734464Z","iopub.status.idle":"2022-09-19T11:10:09.757781Z","shell.execute_reply.started":"2022-09-19T11:10:09.734417Z","shell.execute_reply":"2022-09-19T11:10:09.756785Z"},"trusted":true},"execution_count":null,"outputs":[]}]}