Table 17 The performance for mini-batch K-means algorithm for the last situation Preprocessing 3.
From: Open source Arabic research paper dataset for natural language processing
# K | Silhouette coefficient (euclidean) | Silhouette coefficient (cosine) | ARI | Davies–Bouldin Index |
|---|---|---|---|---|
2 | 0.023 | 0.042 | 0.109 | 5.633 |
3 | 0.048 | 0.086 | 0.095 | 3.573 |
4 | 0.024 | 0.041 | 0.061 | 2.179 |
5 | 0.028 | 0.050 | 0.154 | 3.549 |
6 | 0.025 | 0.044 | 0.171 | 3.307 |
7 | 0.009 | 0.013 | 0.286 | 3.823 |
8 | 0.010 | 0.014 | 0.205 | 3.072 |
9 | 0.006 | 0.007 | 0.221 | 3.259 |
10 | 0.002 | 0.001 | 0.434 | 3.445 |
