CMPT 459 Study Guide - Midterm Guide: Hierarchical Clustering, Bayes Classifier, Laplacian Smoothing

• School: Simon Fraser University
• Department: Computing Science
• Course: CMPT 459
• Professor: Martin Ester

Document Summary

Select your answers for the following questions with a big x. When marking this problem, each of the four possible answers will be considered as a true/false question, and you will receive 1 mark if you correctly marked or correctly not marked that answer: knowledge discovered in a. Natural clusters? that is robust: which is a normalization method to outliers, a k-nearest neighbour classifier . classification discretization visualization data cleaning. Can approximate very complex decision surfaces low test. Consider a 1-dimensional data set {1, 2, 3, 5, 7, 9, 20, 25, 30, 35} and apply single-link. Ties are cases where multiple pairs of clusters have the same (minimum) distance. {1} {2} {3} {5} {7} {9} {20} {25} {30} {35} (b) cut the dendrogram resulting from (a) so that you obtain 3 clusters. {1, 2, 3, 5, 7, 9} {20, 25} {30, 35} (c) the dataset contains three natural (ground truth) clusters c1={1, 2, 3}, c2={5, 7,