Moving script to github

main.py

@@ -0,0 +1,89 @@
+#!/bin/python
+
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+import xgboost as xgb
+from sklearn import preprocessing
+from sklearn.cross_validation import KFold
+from sklearn.ensemble import RandomForestRegressor
+
+train = pd.read_csv('data/Train.csv')
+test = pd.read_csv('data/Test.csv')
+
+#train['Item_Outlet_Sales'] = np.log(train['Item_Outlet_Sales'])
+train= train.fillna(0)
+test = test.fillna(0)
+cat_var = ['Item_Fat_Content','Item_Type','Outlet_Size','Outlet_Location_Type','Outlet_Type']
+def dummy(col,data):
+    for a in col:
+        dummy_train = pd.get_dummies(data[a],prefix=a)
+        data = pd.concat([data,dummy_train],axis=1)
+        data.drop(a,axis=1,inplace=True)
+    return data
+train = dummy(cat_var,train)
+test = dummy(cat_var,test)
+train.to_csv('data/train-cleaned.csv',index=False)
+test.to_csv('data/test-cleaned.csv',index=False)
+
+train = pd.read_csv('data/train-cleaned.csv')
+test = pd.read_csv('data/test-cleaned.csv')
+
+train['Outlet_Identifier_2'] = pd.factorize(train['Outlet_Identifier'])[0]
+
+test['Outlet_Identifier_2'] = pd.factorize(test['Outlet_Identifier'])[0]
+
+cat = ['Item_Identifier']
+
+for a in cat:
+    lb = preprocessing.LabelEncoder()
+    full_var_data = pd.concat((train[a],test[a]),axis=0).astype('str')
+    lb.fit( full_var_data )
+    train[a] = lb.transform(train[a])
+    test[a] = lb.transform(test[a])
+
+def evaluation(x,y):
+    sum =0
+    for a,b in zip(x,y):
+        sum = sum + np.square(a-b) 
+    return np.sqrt(sum/len(x))
+
+from sklearn.metrics import roc_auc_score
+from sklearn.metrics import mean_squared_error
+
+
+train_y = np.array(train["Item_Outlet_Sales"])
+
+## Creating the IDVs from the train and test dataframe ##
+train_X = train.copy()
+test_X = test.copy()
+
+train_X = np.array( train_X.drop(['Item_Outlet_Sales','Outlet_Identifier'],axis=1) )
+
+kfolds = KFold(train_X.shape[0], n_folds=6)
+for dev_index, val_index in kfolds:
+    dev_X, val_X = train_X[dev_index,:], train_X[val_index,:]
+    dev_y, val_y = train_y[dev_index], train_y[val_index]
+
+    reg = RandomForestRegressor(n_estimators=70, max_depth=6, min_samples_leaf=1, max_features="auto", n_jobs=4, random_state=88888)
+    reg.fit(dev_X, dev_y)
+    pred_val_y = reg.predict(val_X)
+    '''dtrain = xgb.DMatrix(dev_X,label = dev_y)
+    dtest = xgb.DMatrix(val_X)
+    bst = xgb.train( plst,dtrain, num_rounds)
+    ypred = bst.predict(dtest,ntree_limit=bst.best_iteration)
+    pred_val_y = (pred_val_y + ypred) / 2'''
+    print np.sqrt(mean_squared_error(val_y, pred_val_y))
+
+print "Building RF1"
+reg = RandomForestRegressor(n_estimators=70, max_depth=6, min_samples_leaf=1, max_features="auto", n_jobs=4, random_state=88888)
+reg.fit(train_X, train_y)
+pred = reg.predict(test_X.drop(['Outlet_Identifier'],axis=1))
+
+test['Item_Identifier'] = lb.inverse_transform(test['Item_Identifier'])
+
+test['Item_Outlet_Sales'] = pred
+
+#test['Item_Outlet_Sales'] =  np.exp(test['Item_Outlet_Sales'])
+
+test.to_csv('submission/sumbit-8.csv',columns=['Item_Identifier','Outlet_Identifier','Item_Outlet_Sales'],index=False)