[Retracted] The Mongolian Vowel Acoustic Model Based on the Clustering Algorithm

Table 8

Recognition results of decision tree state bundling.

Subjects	Sentence level recognition experimental results				Word level recognition experimental results
	% Correct	H	S	N	% Correct	Acc	H	D	S	I	N
Mongol 1	16.57	145	730	875	84.85	80.49	15206	422	2292	783	17920
Mongol 2	28.25	200	508	708	90.7	86.1	10225	179	870	518	11274
Mongol-all	21.79	345	1238	1583	87.11	82.65	25431	601	3162	1301	29194
Dialogue 1	43.23	1855	2435	4290	83.2	73.74	16117	461	2793	1833	19371
Dialogue 2	44.84	1538	1891	3429	82.71	74.37	12349	293	2289	1245	14931
Dialogue 3	37.95	1020	1668	2688	80.72	71.99	11306	278	2423	1222	14007
Dialogue-all	42.4	4413	5994	10407	82.33	73.43	39772	1032	7505	4300	48309