From: From free text to clusters of content in health records: an unsupervised graph partitioning approach
 | Similarity to hand-coded categories (NMI) | Topic Coherence (\(\widehat {PMI}\)) | ||
---|---|---|---|---|
No. of topics/clusters | LDA | MS | LDA | MS |
3 | 0.311 | 0.267 | 2.991 | 3.033 |
7 | 0.409 | 0.393 | 3.218 | 3.303 |
12 | 0.361 | 0.398 | 3.270 | 3.517 |
17 | 0.390 | 0.401 | 3.419 | 3.457 |
44 | 0.395 | 0.388 | 3.549 | 3.716 |