Question

Question I (20 pts): (Regression) Load the Boston dataset from sklearn. Split the dataset into training and testing parts. Use ridge regression (linear_model.Ridge) to predict the target values. Try with both the original attributes and polynomial features (preprocessing.PolynomialFeatures(2, interaction_only=True)). Determine the best regularization coefficient (alpha) in each case. Plot true value vs. predicted values. Finally, decide on a single attribute, try to predict the values using a single attribute. Question II (20 pts): (Classification) Load the MNIST dataset from sklearn. Split the dataset into training and testing parts. Use the KNN classifier. Try different K values and analyze the performances. Write your analyzes as a comment in the code.

          Question I (20 pts):
(Regression) Load the Boston dataset from sklearn. Split the dataset into training and testing
parts. Use ridge regression (linear_model.Ridge) to predict the target values. Try with both the original
attributes and polynomial features (preprocessing.PolynomialFeatures(2, interaction_only=True)).
Determine the best regularization coefficient (alpha) in each case. Plot true value vs. predicted values.
Finally, decide on a single attribute, try to predict the values using a single attribute.
Question II (20 pts):
(Classification) Load the MNIST dataset from sklearn. Split the dataset into training and testing
parts. Use the KNN classifier. Try different K values and analyze the performances. Write your analyzes
as a comment in the code.

Question I (20 pts):
(Regression) Load the Boston dataset from sklearn. Split the dataset into training and testing
parts. Use ridge regression (linearmodel.Ridge) to predict the target values. Try with both the original
attributes and polynomial features (preprocessing.PolynomialFeatures(2, interactiononly=True)).
Determine the best regularization coefficient (alpha) in each case. Plot true value vs. predicted values.
Finally, decide on a single attribute, try to predict the values using a single attribute.
Question II (20 pts):
(Classification) Load the MNIST dataset from sklearn. Split the dataset into training and testing
parts. Use the KNN classifier. Try different K values and analyze the performances. Write your analyzes
as a comment in the code.

Added by Jaime H.

Question

Please give Ace some feedback