Research Article
An Improved Speech Segmentation and Clustering Algorithm Based on SOM and K-Means
Table 5
Comparison of speaker clustering results based on three algorithms.
| ā | K-means speaker clustering (%) | SOM speaker clustering (%) | Improved k-means speaker clustering based on SOM (%) |
| Sound recording 1 | 84.5 | 88.5 | 94.8 |
| Sound recording 2 | 84.3 | 89.2 | 95.1 |
| Sound recording 3 | 85.2 | 83.7 | 94.9 |
| Sound recording 4 | 86.0 | 90.2 | 96.1 |
| Sound recording 5 | 82.2 | 85.5 | 93.6 |
| Sound recording 6 | 81.0 | 82.2 | 90.2 |
| Sound recording 7 | 82.0 | 81.5 | 89.8 |
| Sound recording 8 | 74.8 | 77.2 | 85.5 |
| Sound recording 9 | 73.8 | 76.8 | 86.0 |
| Sound recording 10 | 73.3 | 78.0 | 84.8 |
|
|