HW 10.docx - Purpose Tasks 1 2 ,tw

This preview shows page 1 - 16 out of 16 pages.

Purpose : Evaluate the relevance of different input parameters over the quality of outlier detection. Tasks: 1. Use the workflow provided for this lecture and try to vary some parameters of the hierarchical clustering technique for outlier detection. 2. Use two values for number of cluster, two values for linkage type, and two  distance metrics to determine the number of outliers in the given data. 3. Contrast and compare the results of the experiment using the Cohen’s kappa  measure from the “Scorer” node. Submit your results to this assignment.
Scatter Plot
Scatter Matrix
Hierarchical Clustering Number output cluster: 7 Distance Function: Euclidean Linkage type: Single
Number output cluster: 7 Distance Function: Euclidean

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture