Commit Live PR

q01_pipeline/build.py

@@ -1,13 +1,49 @@
+# %load q01_pipeline/build.py
 import pandas as pd
 import numpy as np
 from sklearn.model_selection import train_test_split, GridSearchCV
 from sklearn.preprocessing import LabelEncoder
 from sklearn.ensemble import RandomForestClassifier
 from sklearn.utils.class_weight import compute_class_weight
 from sklearn.metrics import roc_auc_score
+from imblearn.over_sampling import SMOTE
 
 bank = pd.read_csv('data/Bank_data_to_class.csv', sep=',')
+y = bank['y']
+X = bank.drop(['y'], axis=1)
+
+columns_to_encode = ['job','marital','education','default','housing','loan','contact','day','month','poutcome']
+lbe = LabelEncoder()
+for column  in columns_to_encode:
+    X[column] = lbe.fit_transform(X[column])
+y = lbe.fit_transform(y)
+X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.3,random_state=9)
+model = [RandomForestClassifier()]
+
 
 # Write your solution here :
+def pipeline(X_train1,X_test1,y_trai1n,y_test1,model1):
+    bank = pd.read_csv('data/Bank_data_to_class.csv', sep=',')
+    y = bank['y']
+    X = bank.drop(['y'], axis=1)
+    columns_to_encode = ['job','marital','education','default','housing','loan','contact','day','month','poutcome']
+    lbe = LabelEncoder()
+    for column  in columns_to_encode:
+        X[column] = lbe.fit_transform(X[column])
+    y = lbe.fit_transform(y)
+    X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.3,random_state=9)
+    smote = SMOTE(random_state=9, kind='borderline2')
+    X_sampled_train,y_sampled_train = smote.fit_sample(X_train,y_train)
+    rfc = RandomForestClassifier(class_weight={0:1,1:2},random_state=9)
+    param_grid={
+        'criterion' : ['gini','entropy'],
+        'n_estimators' : [50,75,100],
+        'max_depth' : [6,8,10],
+        'min_samples_leaf' : [10,15,20]
+    }
+    grid_search =GridSearchCV(estimator=rfc,param_grid=param_grid,cv=5)
+    grid_search.fit(X_sampled_train,y_sampled_train)
+    y_pred = grid_search.predict(X_test)
+    return grid_search,roc_auc_score(y_test,y_pred)