Research Article
Acoustic Model with Multiple Lexicon Types for Indonesian Speech Recognition
Table 9
The performance of the acoustic models using 10,000 unvalidated utterances.
| | Lexicon type | GMM-HMM model (%WER) | | Monophone | Triphone DELTA + DELTA-DELTA | Triphone LDA + MLLT | Triphone SAT |
| | 1c-vocab-full | 85.61 | 85.26 | 71.30 | 71.24 | | enmap-1c-vocab-full | 86.83 | 86.12 | 73.06 | 73.07 | | enmap-kv-vocab-full | 85.30 | 91.16 | 84.28 | 84.06 | | kv-vocab-full | 84.12 | 91.11 | 85.28 | 85.69 |
|
|
