16. “K-Means++ Algorithm”
2. 클러스터의 중심점
1. Random initial centroid
2. Calculate distance, D(x)
3. Choose next centroid from D(x)2
“2007. k-means++: The Advantages of Careful Seeding” 논문 참조
21. 4. 클러스터링 검증방법
1. 내부 평가 (Internal Evaluation)
Davies-Bouldin Index, Dunn Index
2. 외부 평가 (External Evaluation)
Rand Measure, F-Measure, Jaccard Index