Acoustic Model with Multiple Lexicon Types for Indonesian Speech Recognition

Table 8

The performance of the acoustic models using 10,000 validated utterances.

Lexicon type	GMM-HMM model (%WER)
	Monophone	Triphone DELTA + DELTA-DELTA	Triphone LDA + MLLT	Triphone SAT
1c-vocab-full	82.30	79.61	62.80	63.20
enmap-1c-vocab-full	82.47	81.05	64.25	64.54
enmap-kv-vocab-full	80.93	86.86	76.61	75.33
kv-vocab-full	79.27	88.00	77.36	76.47