An alternative to recursive partitioning is the k th nearest neighbor method. This method computes the distance between 2 messages using the values of the derived variables from Section 3.9, or some subset of them. Use the email in train DF to find the k closest messages to each email in test DF. Then use these k neighbors to classify the test message as spam or ham. That is, use the neighbors' classifications to vote on the classi- fication for the test message. Compare this method for predicting spam to the recursive partitioning approach. Use both Type I and Type II errors in making your comparison. Include a comparison of the two approaches from a computational perspective. Is one much faster or easier to implement?
