Homework – KNN

Paper, Order, or Assignment Requirements

 

Homework 2 is based on problem 7.1 on page 146. Instead of doing the original problems in the textbook, you should do problems in this document. The data set involved is UniversalBank.xls, which is included in the zip file posted in the Assignments area. 

Step 1: Preprocess the  data set by converting categorical variable “Education” into binary dummy variables. The new data set is called transformed data hereafter. 

Step 2: Partition the transformed data into training (60%) and validation (40%) sets, using the default random seed 12345. (There are two partition icons in XLMiner. The  data mining partition should be used.)

Step 3: Perform “k-nearest neighbors” analysis using all predicators  except ID and ZIP code with the goal of finding the best k. Specify the success class as 1 (loan acceptance), and the default cutoff value of 0.5. Personal Loan should be the output variable.