Issue 18/08/2011 Revision Date 11/06/2013 Form No. ECT/AC/F.05.02
Case Study on KNN Classification
Date Time: Total Mark:
10
Student’s Name
Student’s ID
Course Name Business Intelligence
Course Code BIT405
Semester Spring 2017
Instructor’s Name Dr. Myriam Bounhas, Dr. Bilel Elayeb, Dr. Wala Saber
Scope and Focus:
● Classification Methods
● The Naive Rule
● Naive Bayes
● k-Nearest Neighbors
Contributing to the following CLOs:
CLO #1 Describe the role of Business Intelligence in an organization.
CLO #2 Understand the data mining process and its related issues.
CLO #3 Create, evaluate and apply different intelligence models.
Case study: KNN classification
Consider the following data concerning credit default (cf. Figure 1). Age and Loan are two numerical variables (predictors) and Default is the target.
Figure 1: Loan vs. Age
1). Regular Distance
(a) We can now use the training set to classify an unknown case (Age=48 and Loan=$142,000) using Euclidean distance. Complete the calculus of distances by filling Table 1.
Table 1: Euclidean distance calculus
(b) Determine the prediction for the unknown case if K= 1. Provide the detail of calculus for the first nearest neighbor.
.............................................................................................................................................................
.............................................................................................................................................................
.............................................................................................................................................................
.............................................................................................................................................................
.............................................................................................................................................................
.............................................................................................................................................................
(c) Determine the prediction for the unknown case if K= 3. Provide the detail of calculus for each nearest neighbor.
.............................................................................................................................................................
.............................................................................................................................................................
.............................................................................................................................................................
.............................................................................................................................................................
.............................................................................................................................................................
............................................................................................................................................................. .............................................................................................................................................................
.............................................................................................................................................................
.............................................................................................................................................................
.............................................................................................................................................................