final version

bias_varience_decomp.py

@@ -1,4 +1,5 @@
 import numpy as np
+import pandas as pd
 
 def _draw_bootstrap_sample(rng, X, y, frac = 1.0):
     """
@@ -36,6 +37,15 @@ def bias_variance_decomp(estimator, X_train, y_train, X_test, y_test,
 
     mean_predictions = np.mean(all_pred, axis=0)
 
-    avg_bias = np.sum((mean_predictions - y_test)**2) / y_test.size
+    avg_bias = np.mean((mean_predictions - y_test)**2)
     avg_var = np.sum((mean_predictions - all_pred)**2) / all_pred.size
-    return avg_expected_loss, avg_bias, avg_var
+    return avg_expected_loss, avg_bias, avg_var
+
+def bias_variance_plot(alphas, metrics, figsize):
+
+    complexity = np.log(alphas)
+    df = pd.DataFrame(metrics, columns=['mse', 'bias', 'varience'])
+    df['complexity'] = complexity
+    df = df.set_index('complexity')
+
+    df.plot(figsize=figsize)

utils.py

@@ -1,6 +1,30 @@
+import numpy as np
+from sklearn.metrics import mean_squared_error, r2_score
+
 def generate_poly_features(df, cols, max_degree=3):
+    """
+    Function to add polynomial transformation (x -> x^i) for column names in cols (iterable) in df (Dataframe)
+    max_degree - maximum degree of polynomial features to be added
+    """
     df = df.copy()
     for col in cols:
-        for degree in range(max_degree):
+        for degree in range(1, max_degree+1):
             df[f"{col}^{degree}"] = df[col]**degree
     return df 
+
+def summary_model(estimator, summary_name, X, y):
+
+    y_pred = estimator.predict(X)
+    r2 = r2_score(y, y_pred)
+    rmse = np.sqrt(mean_squared_error(y, y_pred))
+
+    print(f"The model performance for {summary_name}")
+    print("--------------------------------------")
+    print('RMSE is {}'.format(rmse))
+    print('R2 score is {}'.format(r2))
+    print("\n")
+
+
+
+
+