H4 - Statistical Pattern Recognition University of...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Statistical Pattern Recognition University of Maryland, College Park Professor Rama Chellappa October 1, 2009 Handout 4 Homework 4 This problem set is due Thursday October 8 at 11:00AM. 1. Consider two classes ω1 , ω2 in the two-dimensional space. The data from class ω1 are uniformly distributed inside a circle of radius r . The data of class ω2 are also uniformly distributed inside another circle of radius r . The distance between the centers of the circles is greater than 4r . Let N be the number of available training samples. Show that the probability of error of the NN classifier is always smaller than that of the k NN, for any k ≥ 3. 2. Generate 50 feature vectors for each of the two classes p(x|ω1 ) ∼ N p(x|ω1 ) ∼ N 1 1 , σ2I , 1.5 , σ 2 I , σ 2 = 0.2 and use them as training points. In the sequel, 1.5 generate 100 vectors from each class and classify them according to the NN and 3NN rules. Compute the classification error percentages. ...
View Full Document

This note was uploaded on 10/26/2009 for the course CMSC 828 taught by Professor Staff during the Fall '05 term at Maryland.

Ask a homework question - tutors are online