Table 11 The performance for mini-batch K-means algorithm for the first situation used the text files with Preprocessing 1.
From: Open source Arabic research paper dataset for natural language processing
# K | Silhouette coefficient (euclidean) | Silhouette coefficient (cosine) | ARI | Davies–Bouldin Index |
|---|---|---|---|---|
2 | 0.013 | 0.022 | 0.004 | 4.827 |
3 | 0.028 | 0.050 | 0.037 | 2.597 |
4 | 0.007 | 0.011 | 0.040 | 2.359 |
5 | 0.030 | 0.054 | 0.125 | 2.922 |
6 | 0.031 | 0.056 | 0.120 | 3.038 |
7 | 0.022 | 0.038 | 0.101 | 2.615 |
8 | −0.001 | -0.005 | 0.147 | 3.348 |
9 | 0.004 | 0.004 | 0.120 | 3.607 |
10 | −0.019 | -0.039 | 0.122 | 2.952 |
