Bayesian classifier. Can you think of ways that the Bayesian classifier you built in Chapter 6 could be used on the matchmaker dataset? What would be good examples of features?

Optimizing a dividing line. Do you think it’s possible to choose a dividing line using the optimization methods you learned in Chapter 5, instead of just using the averages? What cost function would you use?

Choosing the best kernel parameters. Write a function that loops over different values for gamma and determines what the best value for a given dataset is.

Hierarchy of interests. Design a simple hierarchy of interests, along with a data structure to represent it. Alter the matchcount function to use the hierarchy to give partial points for matches.