_dsc ASSIGNMENT 5.docx - ASSIGNMENT-5 PROBLEM 1 1 A The...

This preview shows page 1 - 10 out of 12 pages.

ASSIGNMENT-5 PROBLEM 1 1. A The squared Euclid distance was determined in order to classify cluster centers that have the shortest square distance between each data-point. B To measure similarity, the square Euclidean distance was used. C a. I. K= 3 II. K=4 III. K=5
IV. K=6 b. I. K=3 II. K=4
III. K=5 IV. K=6 C I. K=3
II. K=4 III. K=5 IV. K=6 D K=3 gives the best split as the maximum precision figure is 89.5%. E
The accuracy rate improved to 91.9 percent when the Z raters of the attributes were calculated, and the K-means algorithm regenerated. 2. A. 1)
2) Class Count 1 202 2 6 3 2 B. 1)
2) 3. Full linkage performs better on this dataset than single linkage on the basis of the above performance, but K-means classification performs much better than the hierarchical clustering approach of total linkage. 4. This article discusses kernel groupings belonging to three distinct varieties of wheat: Kama, Rosa and Canadian. For each wheat type, seventy data points are provided. To

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture