commit-live-students · rameshgayam · Nov 10, 2017
diff --git a/__init__.pyc b/__init__.pyc
diff --git a/q01_pipeline/__init__.pyc b/q01_pipeline/__init__.pyc
diff --git a/q01_pipeline/build.py b/q01_pipeline/build.py
@@ -1,3 +1,4 @@
+# %load q01_pipeline/build.py
 import pandas as pd
 import numpy as np
 from sklearn.model_selection import train_test_split, GridSearchCV
@@ -6,8 +7,28 @@
 from sklearn.utils.class_weight import compute_class_weight
 from sklearn.metrics import roc_auc_score
 
+
+
 bank = pd.read_csv('data/Bank_data_to_class.csv', sep=',')
 
-# Write your solution here :
+#Label encode
+le = LabelEncoder()
+for col in bank.select_dtypes(include=["object"]).columns.values:
+    bank[col] = le.fit_transform(bank[col])
+
+param_grid = {"n_estimators": [10, 50, 120],
+              "max_depth": [40, 20, 10],
+              "max_leaf_nodes": [5, 10, 2]}
+X = bank.iloc[:,:-1]
+y = bank['y']
+X_train,X_test,y_train,y_test = train_test_split(X,y,test_size = 0.3, random_state=9)
+
+model = RandomForestClassifier(random_state=9,class_weight='balanced')
+
 
+def pipeline(X_train,X_test,y_train,y_test,model):
 
+    gscv = GridSearchCV(model, param_grid,cv=5)
+    gscv.fit(X_train,y_train)
+    y_pred = gscv.predict(X_test)
+    return gscv, roc_auc_score(y_test,y_pred).item()
diff --git a/q01_pipeline/build.pyc b/q01_pipeline/build.pyc
diff --git a/q01_pipeline/tests/__init__.pyc b/q01_pipeline/tests/__init__.pyc
diff --git a/q01_pipeline/tests/test_q01_pipeline.pyc b/q01_pipeline/tests/test_q01_pipeline.pyc