commit-live-students · sannidh · Nov 25, 2018 · Nov 25, 2018 · Nov 25, 2018 · Nov 25, 2018
diff --git a/__pycache__/__init__.cpython-36.pyc b/__pycache__/__init__.cpython-36.pyc
diff --git a/q01_outlier_removal/__pycache__/__init__.cpython-36.pyc b/q01_outlier_removal/__pycache__/__init__.cpython-36.pyc
diff --git a/q01_outlier_removal/__pycache__/build.cpython-36.pyc b/q01_outlier_removal/__pycache__/build.cpython-36.pyc
diff --git a/q01_outlier_removal/build.py b/q01_outlier_removal/build.py
@@ -1,8 +1,19 @@
+# %load q01_outlier_removal/build.py
 # Default imports
+from greyatomlib.logistic_regression_project.q01_outlier_removal.build import outlier_removal
 import pandas as pd
 
 loan_data = pd.read_csv('data/loan_prediction_uncleaned.csv')
 loan_data = loan_data.drop('Loan_ID', 1)
 
+def outlier_removal(df):
+    qv = 0.95
+    numeric_features = ['ApplicantIncome', 'CoapplicantIncome', 'LoanAmount']
+    df_qv = df.quantile(q=qv, axis=0, numeric_only=True, interpolation='linear')
+    for feature in numeric_features:
+        df = df.drop(df[df[feature] > df_qv[feature]].index)
+    return df
+
+
+
 
-# Write your Solution here:
diff --git a/q01_outlier_removal/tests/__pycache__/__init__.cpython-36.pyc b/q01_outlier_removal/tests/__pycache__/__init__.cpython-36.pyc
diff --git a/q01_outlier_removal/tests/__pycache__/test_q01_outlier_removal.cpython-36.pyc b/q01_outlier_removal/tests/__pycache__/test_q01_outlier_removal.cpython-36.pyc
diff --git a/q02_data_cleaning_all/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all/__pycache__/__init__.cpython-36.pyc
diff --git a/q02_data_cleaning_all/__pycache__/build.cpython-36.pyc b/q02_data_cleaning_all/__pycache__/build.cpython-36.pyc
diff --git a/q02_data_cleaning_all/build.py b/q02_data_cleaning_all/build.py
@@ -1,14 +1,28 @@
+# %load q02_data_cleaning_all/build.py
 # Default Imports
 import sys, os
 sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname('__file__'))))
 import pandas as pd
 import numpy as np
 from sklearn.model_selection import train_test_split
 from greyatomlib.logistic_regression_project.q01_outlier_removal.build import outlier_removal
-
+from sklearn.preprocessing import Imputer
 loan_data = pd.read_csv('data/loan_prediction_uncleaned.csv')
 loan_data = loan_data.drop('Loan_ID', 1)
 loan_data = outlier_removal(loan_data)
 
+def data_cleaning(df):
+    imputer_mean = Imputer(missing_values='NaN', strategy='mean')
+    imputer_mean.fit(df[['LoanAmount']])
+    df['LoanAmount'] = imputer_mean.transform(df[['LoanAmount']])
+    cat_features = ['Gender', 'Married', 'Dependents', 'Self_Employed', 'Loan_Amount_Term', 'Credit_History']
+    for feature in cat_features:
+        df[feature] = df[feature].fillna(df[feature].mode()[0])
+    y = df.iloc[:,-1]
+    X = df.iloc[:,:-1]
+    np.random.seed(9)
+    X_train, X_test, y_train, y_test = train_test_split(X,y, test_size=0.25, train_size=0.75)
+    return X, y, X_train, X_test, y_train, y_test
+
+
 
-# Write your solution here :
diff --git a/q02_data_cleaning_all/tests/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all/tests/__pycache__/__init__.cpython-36.pyc
diff --git a/q02_data_cleaning_all/tests/__pycache__/test_q02_data_cleaning.cpython-36.pyc b/q02_data_cleaning_all/tests/__pycache__/test_q02_data_cleaning.cpython-36.pyc
diff --git a/q02_data_cleaning_all_2/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all_2/__pycache__/__init__.cpython-36.pyc
diff --git a/q02_data_cleaning_all_2/__pycache__/build.cpython-36.pyc b/q02_data_cleaning_all_2/__pycache__/build.cpython-36.pyc
diff --git a/q02_data_cleaning_all_2/build.py b/q02_data_cleaning_all_2/build.py
@@ -1,3 +1,4 @@
+# %load q02_data_cleaning_all_2/build.py
 # Default Imports
 import pandas as pd
 import numpy as np
@@ -9,5 +10,19 @@
 loan_data = outlier_removal(loan_data)
 X, y, X_train, X_test, y_train, y_test = data_cleaning(loan_data)
 
+def data_cleaning_2(X_train, X_test, y_train, y_test):
+    numerical_features = ['ApplicantIncome', 'CoapplicantIncome', 'LoanAmount', 'Loan_Amount_Term','Credit_History']
+    categorical_features = ['Gender', 'Married', 'Dependents', 'Education', 'Self_Employed', 'Property_Area']
+    for feature in numerical_features:
+        X_train[feature] = np.sqrt(X_train[feature])
+        X_test[feature] = np.sqrt(X_test[feature])
+    X_train_dummy = pd.get_dummies(X_train[categorical_features], drop_first=True)
+    X_test_dummy = pd.get_dummies(X_test[categorical_features], drop_first=True)
+    X_train = X_train[numerical_features].join(X_train_dummy)
+    X_test = X_test[numerical_features].join(X_test_dummy)
+    y_train, y_test = y_train, y_test
+    return X_train, X_test, y_train, y_test
+
+
+
 
-# Write your solution here :
diff --git a/q02_data_cleaning_all_2/tests/__pycache__/__init__.cpython-36.pyc b/q02_data_cleaning_all_2/tests/__pycache__/__init__.cpython-36.pyc
diff --git a/q02_data_cleaning_all_2/tests/__pycache__/q02_test_data_cleaning_2.cpython-36.pyc b/q02_data_cleaning_all_2/tests/__pycache__/q02_test_data_cleaning_2.cpython-36.pyc
diff --git a/q03_logistic_regression/__pycache__/__init__.cpython-36.pyc b/q03_logistic_regression/__pycache__/__init__.cpython-36.pyc
diff --git a/q03_logistic_regression/__pycache__/build.cpython-36.pyc b/q03_logistic_regression/__pycache__/build.cpython-36.pyc
diff --git a/q03_logistic_regression/build.py b/q03_logistic_regression/build.py
@@ -1,3 +1,4 @@
+# %load q03_logistic_regression/build.py
 # Default Imports
 import pandas as pd
 from sklearn.preprocessing import StandardScaler
@@ -6,13 +7,27 @@
 from greyatomlib.logistic_regression_project.q01_outlier_removal.build import outlier_removal
 from greyatomlib.logistic_regression_project.q02_data_cleaning_all.build import data_cleaning
 from greyatomlib.logistic_regression_project.q02_data_cleaning_all_2.build import data_cleaning_2
-
+from sklearn.metrics import mean_squared_error, r2_score, accuracy_score, confusion_matrix, f1_score, log_loss
+from sklearn.linear_model import LinearRegression, Lasso, Ridge, ElasticNet, LogisticRegression
+from sklearn.model_selection import train_test_split
+from sklearn.preprocessing import PolynomialFeatures, StandardScaler, LabelEncoder
 loan_data = pd.read_csv('data/loan_prediction_uncleaned.csv')
 loan_data = loan_data.drop('Loan_ID', 1)
 loan_data = outlier_removal(loan_data)
 X, y, X_train, X_test, y_train, y_test = data_cleaning(loan_data)
 X_train, X_test, y_train, y_test = data_cleaning_2(X_train, X_test, y_train, y_test)
 
 
-# Write your solution code here:
+def logistic_regression(X_train, X_test, y_train, y_test):
+    model = LogisticRegression(random_state=9)
+    scaling_features =['ApplicantIncome', 'CoapplicantIncome', 'LoanAmount']
+    scaler = StandardScaler()
+    X_train[scaling_features] = scaler.fit_transform( X_train[scaling_features])
+    X_test[scaling_features] = scaler.fit_transform( X_test[scaling_features])
+    model.fit(X_train, y_train)
+    y_pred = model.predict(X_test)
+    cm = confusion_matrix(y_test,y_pred)
+    return cm
+
+
 
diff --git a/q03_logistic_regression/tests/__pycache__/__init__.cpython-36.pyc b/q03_logistic_regression/tests/__pycache__/__init__.cpython-36.pyc
diff --git a/q03_logistic_regression/tests/__pycache__/test_q03_logistic_regression.cpython-36.pyc b/q03_logistic_regression/tests/__pycache__/test_q03_logistic_regression.cpython-36.pyc