Ii number of elements in each cluster k3 k4 k5 k6 iii

This preview shows page 3 - 10 out of 13 pages.

ii. Number of elements in each cluster K=3 K=4 K=5 K=6 iii. The class distribution within each cluster
DSC441- Fall 2018, Assignment 5, Page 4 of 13 K=3 K=4 K=5 K=6 iv. In your opinion, which k should be selected? Explain your selection.
DSC441- Fall 2018, Assignment 5, Page 5 of 13
v. For the selected k in iv, analyze and report if the normalization of the attributes will influence the clustering results.
ii. (10 points) Perform hierarchical clustering using all attributes except the class label as follows: i. Apply single linkage algorithm and report 1. The dendogram
DSC441- Fall 2018, Assignment 5, Page 6 of 13 Single linkage dendogram:
DSC441- Fall 2018, Assignment 5, Page 7 of 13 2. The class distribution at the level of the dendogram where there are only three clusters. Class Count 1 202 2 6 3 2 ii. Apply complete linkage and report 1. The dendogram
DSC441- Fall 2018, Assignment 5, Page 8 of 13 Complete linkage dendogram
DSC441- Fall 2018, Assignment 5, Page 9 of 13
iii. (2.5 points) Compare the results with hierarchical clustering and k-means algorithm.
iv. (2.5 points) Create an executive summary (~half a page) that outlines the problem, summarizes the data, describes the methodology, summarizes the results, and makes recommendations. When creating it, imagine that you will give this summary to someone who is not an expert in data mining.

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture