Chapter 18 Flashcards
1
Q
Types of distance based clustering covere in the course:
A
1) K-means
2) agglomerative hierachical
2
Q
weaknesses k-means:
A
1) euclidean = prefers round clusters - outliers and scaling canaffect it.
2) Initialization.
3) sometimes clusters contain no points - remove or relocate.
4) requires suitable K choice
3
Q
How to measure a difference in clustering:
A
1) Rand index
2) Jaccard similarity
3) normalized mutual information
