Table 2 Clustering metrics for both datasets via different feature numbers.

From: Enhancing classification accuracy in medical datasets using a hybrid distance and cluster refinement-based K-means clustering method

# Features

Breast Cancer Wisconsin (BCW)

Heart Disease

Acc.

F1

Homo.

ARI

Acc.

F1

Homo.

ARI

5

0.9737

0.9663

0.8532

0.8964

0.7667

0.6957

0.4965

0.5714

10

0.9649

0.9545

0.7849

0.8631

0.90

0.8696

0.5352

0.6334

15-13

0.9649

0.9524

0.7721

0.8629

0.8325

0.7428

0.5169

0.6227

20

0.9825

0.9631

0.8617

0.9303

—

—

—

—

25

0.9825

0.9647

0.8676

0.9303

—

—

—

—

30

0.9737

0.9647

0.8165

0.8963

—

—

—

—

Back to article page