commit-live-students · preetiail · Nov 8, 2018 · Nov 8, 2018 · Nov 8, 2018 · Nov 8, 2018
diff --git a/__pycache__/__init__.cpython-36.pyc b/__pycache__/__init__.cpython-36.pyc
diff --git a/q01_load_data/__pycache__/__init__.cpython-36.pyc b/q01_load_data/__pycache__/__init__.cpython-36.pyc
diff --git a/q01_load_data/__pycache__/build.cpython-36.pyc b/q01_load_data/__pycache__/build.cpython-36.pyc
diff --git a/q01_load_data/build.py b/q01_load_data/build.py
@@ -1,8 +1,11 @@
+# %load q01_load_data/build.py
 # Default imports
 import pandas as pd
 
 path = 'data/house_prices_multivariate.csv'
+def load_data(path):
+    return(pd.read_csv(path))
 
+load_data(path)
 
-# Write your code here :
 
diff --git a/q01_load_data/tests/__pycache__/__init__.cpython-36.pyc b/q01_load_data/tests/__pycache__/__init__.cpython-36.pyc
diff --git a/q01_load_data/tests/__pycache__/test_q01_load_data.cpython-36.pyc b/q01_load_data/tests/__pycache__/test_q01_load_data.cpython-36.pyc
diff --git a/q02_data_splitter/__pycache__/__init__.cpython-36.pyc b/q02_data_splitter/__pycache__/__init__.cpython-36.pyc
diff --git a/q02_data_splitter/__pycache__/build.cpython-36.pyc b/q02_data_splitter/__pycache__/build.cpython-36.pyc
diff --git a/q02_data_splitter/build.py b/q02_data_splitter/build.py
@@ -1,8 +1,12 @@
+# %load q02_data_splitter/build.py
 # Default Imports
 from greyatomlib.linear_regression.q01_load_data.build import load_data
 import pandas as pd
-df = load_data('data/house_prices_multivariate.csv')
-
 
-# Your Code Here
+df = load_data('data/house_prices_multivariate.csv')
+def data_splitter(df):
+    y=df.iloc[:,]
+    x=df.iloc[:,0:34]
+    return(x,y)
+data_splitter(df)
 
diff --git a/q02_data_splitter/tests/__pycache__/__init__.cpython-36.pyc b/q02_data_splitter/tests/__pycache__/__init__.cpython-36.pyc
diff --git a/q02_data_splitter/tests/__pycache__/test_q02_data_splitter.cpython-36.pyc b/q02_data_splitter/tests/__pycache__/test_q02_data_splitter.cpython-36.pyc
diff --git a/q03_linear_regression/__pycache__/__init__.cpython-36.pyc b/q03_linear_regression/__pycache__/__init__.cpython-36.pyc
diff --git a/q03_linear_regression/__pycache__/build.cpython-36.pyc b/q03_linear_regression/__pycache__/build.cpython-36.pyc
diff --git a/q03_linear_regression/build.py b/q03_linear_regression/build.py
@@ -1,10 +1,19 @@
+# %load q03_linear_regression/build.py
 from greyatomlib.linear_regression.q01_load_data.build import load_data
 from greyatomlib.linear_regression.q02_data_splitter.build import data_splitter
 from sklearn.linear_model import LinearRegression
+#import matplotlib
+#import matplotlib.pyplot as plt
 
 dataframe = load_data('data/house_prices_multivariate.csv')
 X, y = data_splitter(dataframe)
+def linear_regression(X,y):
+    regressor = LinearRegression()
+    lm=regressor.fit(X, y)
+    return(lm)
+
+
+
 
-# Write your code here :
 
 
diff --git a/q03_linear_regression/tests/__pycache__/__init__.cpython-36.pyc b/q03_linear_regression/tests/__pycache__/__init__.cpython-36.pyc
diff --git a/q03_linear_regression/tests/__pycache__/test_q03_linear_regression.cpython-36.pyc b/q03_linear_regression/tests/__pycache__/test_q03_linear_regression.cpython-36.pyc
diff --git a/q04_linear_predictor/__pycache__/__init__.cpython-36.pyc b/q04_linear_predictor/__pycache__/__init__.cpython-36.pyc
diff --git a/q04_linear_predictor/__pycache__/build.cpython-36.pyc b/q04_linear_predictor/__pycache__/build.cpython-36.pyc
diff --git a/q04_linear_predictor/build.py b/q04_linear_predictor/build.py
@@ -1,14 +1,25 @@
+# %load q04_linear_predictor/build.py
 # Default Imports
 from greyatomlib.linear_regression.q01_load_data.build import load_data
 from greyatomlib.linear_regression.q02_data_splitter.build import data_splitter
 from greyatomlib.linear_regression.q03_linear_regression.build import linear_regression
 from sklearn.linear_model import LinearRegression
 from sklearn.metrics import mean_squared_error, mean_absolute_error, r2_score
-
+import matplotlib.pyplot as plt
+import pandas as pd
+from sklearn.metrics import mean_absolute_error
+from sklearn.metrics import r2_score
 
 dataframe = load_data('data/house_prices_multivariate.csv')
 X, y = data_splitter(dataframe)
-linear_model = linear_regression(X, y)
+regressor = linear_regression(X, y)
+
+def linear_predictor(regressor,dataframe,y ): 
+    y_pred = regressor.predict(X)
+    dataresult=pd.DataFrame({'actual': y,'predicted': y_pred})
+    MSE=mean_squared_error(y_pred, y)
+    MAE=mean_absolute_error(y,y_pred)
+    RSE=r2_score(y,y_pred)
+    return(y_pred,MSE, MAE,RSE)
 
 
-# Your code here
diff --git a/q04_linear_predictor/tests/__pycache__/__init__.cpython-36.pyc b/q04_linear_predictor/tests/__pycache__/__init__.cpython-36.pyc
diff --git a/q04_linear_predictor/tests/__pycache__/test_q04_linear_predictor.cpython-36.pyc b/q04_linear_predictor/tests/__pycache__/test_q04_linear_predictor.cpython-36.pyc
diff --git a/q05_residuals/__pycache__/__init__.cpython-36.pyc b/q05_residuals/__pycache__/__init__.cpython-36.pyc
diff --git a/q05_residuals/__pycache__/build.cpython-36.pyc b/q05_residuals/__pycache__/build.cpython-36.pyc
diff --git a/q05_residuals/build.py b/q05_residuals/build.py
@@ -1,14 +1,21 @@
+# %load q05_residuals/build.py
 # Default Imports
 from greyatomlib.linear_regression.q01_load_data.build import load_data
 from greyatomlib.linear_regression.q02_data_splitter.build import data_splitter
 from greyatomlib.linear_regression.q03_linear_regression.build import linear_regression
 from greyatomlib.linear_regression.q04_linear_predictor.build import linear_predictor
 from sklearn.linear_model import LinearRegression
+import numpy as np
+import pandas as pd
+from sklearn.metrics import r2_score
 
 dataframe = load_data('data/house_prices_multivariate.csv')
 X, y = data_splitter(dataframe)
 linear_model = linear_regression(X, y)
 y_pred, _, __, ___ = linear_predictor(linear_model, X, y)
 
+def residuals(y,y_pred):
+    RSE=pd.Series([y[i]-y_pred[i] for i in range(len(dataframe))])
+    return(RSE)
+residuals(y,y_pred)
 
-# Your code here
diff --git a/q05_residuals/tests/__pycache__/__init__.cpython-36.pyc b/q05_residuals/tests/__pycache__/__init__.cpython-36.pyc
diff --git a/q05_residuals/tests/__pycache__/test_q05_residuals.cpython-36.pyc b/q05_residuals/tests/__pycache__/test_q05_residuals.cpython-36.pyc
diff --git a/q06_plot_residuals/__pycache__/__init__.cpython-36.pyc b/q06_plot_residuals/__pycache__/__init__.cpython-36.pyc
diff --git a/q06_plot_residuals/__pycache__/build.cpython-36.pyc b/q06_plot_residuals/__pycache__/build.cpython-36.pyc
diff --git a/q06_plot_residuals/build.py b/q06_plot_residuals/build.py
@@ -1,3 +1,4 @@
+# %load q06_plot_residuals/build.py
 # Default Imports
 from greyatomlib.linear_regression.q01_load_data.build import load_data
 from greyatomlib.linear_regression.q02_data_splitter.build import data_splitter
@@ -6,14 +7,20 @@
 from greyatomlib.linear_regression.q05_residuals.build import residuals
 from sklearn.linear_model import LinearRegression
 import matplotlib.pyplot as plt
-plt.switch_backend('agg')
+#plt.switch_backend('agg')
 
 dataframe = load_data('data/house_prices_multivariate.csv')
 X, y = data_splitter(dataframe)
 linear_model = linear_regression(X, y)
 y_pred, _, __, ___ = linear_predictor(linear_model, X, y)
 error_residuals = residuals(y, y_pred)
+def plot_residuals(y,error_residuals):
+    plt.scatter(y,error_residuals)
+    plt.title('Scatter plot of SalePrice (target) vs Residuals ')
+    plt.xlabel('Target Variable')
+    plt.ylabel('error residuals')
+    plt.show()
+
 
 
-# Your code here
 
diff --git a/q06_plot_residuals/tests/__pycache__/__init__.cpython-36.pyc b/q06_plot_residuals/tests/__pycache__/__init__.cpython-36.pyc
diff --git a/q06_plot_residuals/tests/__pycache__/test_q06_plot_residuals.cpython-36.pyc b/q06_plot_residuals/tests/__pycache__/test_q06_plot_residuals.cpython-36.pyc
diff --git a/q07_hist_residuals/__pycache__/__init__.cpython-36.pyc b/q07_hist_residuals/__pycache__/__init__.cpython-36.pyc
diff --git a/q07_hist_residuals/__pycache__/build.cpython-36.pyc b/q07_hist_residuals/__pycache__/build.cpython-36.pyc
diff --git a/q07_hist_residuals/build.py b/q07_hist_residuals/build.py
@@ -1,3 +1,4 @@
+# %load q07_hist_residuals/build.py
 # Default Imports
 from greyatomlib.linear_regression.q01_load_data.build import load_data
 from greyatomlib.linear_regression.q02_data_splitter.build import data_splitter
@@ -19,3 +20,5 @@
 def hist_residuals(error_residuals, bins=60):
     plt.figure(figsize=(15,8))
     plt.hist(error_residuals, bins=bins)
+
+
diff --git a/q07_hist_residuals/tests/__pycache__/__init__.cpython-36.pyc b/q07_hist_residuals/tests/__pycache__/__init__.cpython-36.pyc
diff --git a/q07_hist_residuals/tests/__pycache__/test_q07_hist_residuals.cpython-36.pyc b/q07_hist_residuals/tests/__pycache__/test_q07_hist_residuals.cpython-36.pyc
diff --git a/q08_qq_residuals/__pycache__/__init__.cpython-36.pyc b/q08_qq_residuals/__pycache__/__init__.cpython-36.pyc
diff --git a/q08_qq_residuals/__pycache__/build.cpython-36.pyc b/q08_qq_residuals/__pycache__/build.cpython-36.pyc
diff --git a/q08_qq_residuals/build.py b/q08_qq_residuals/build.py
@@ -1,3 +1,4 @@
+# %load q08_qq_residuals/build.py
 # Default Imports
 from greyatomlib.linear_regression.q01_load_data.build import load_data
 from greyatomlib.linear_regression.q02_data_splitter.build import data_splitter
@@ -18,5 +19,9 @@
 y_pred, _, __, ___ = linear_predictor(linear_model, X, y)
 error_residuals = residuals(y, y_pred)
 
+def qq_residuals(error_residuals):
+    stats.probplot(error_residuals)
+
+
+
 
-# Your code here
diff --git a/q08_qq_residuals/tests/__pycache__/__init__.cpython-36.pyc b/q08_qq_residuals/tests/__pycache__/__init__.cpython-36.pyc
diff --git a/q08_qq_residuals/tests/__pycache__/test_q08_qq_residuals.cpython-36.pyc b/q08_qq_residuals/tests/__pycache__/test_q08_qq_residuals.cpython-36.pyc