Model Selection Lab Sheet
Preliminaries
The notebook 07 Model Selection Lab shows how to fit a k-NN model for the HotelRevHelpfulness dataset. It assesses three options:
● whether to use a StandardScaler, MinMaxScaler or no scaler.
● what k to use for k-NN
● what weighting policy
Q1
Repeat this model fitting exercise to fit a Naive Bayes mode for the same data. Consider the same scaling options and GaussianNB and BernoulliNB as classifier options.
Q2
Find the best decision tree model for the HotelRevHelpfulness dataset considering max_leaf_nodes and the splitting criterion. The splitting criterion can be either 'gini' or 'entropy', you can select your own options for max_leaf_nodes.
Q3
Which model would you recommend for this dataset?
