Abstract In traditional Katz approach, the discount coefficients may be greater than 1 or can not be calculated in some serious conditions. The idea of smoothing in log domain of couple occurrence number in simple good-turing is adopted. The modified Katz approach is proposed combined with back-off model. The proposed approach is further applied in speech recognition system based on lattice. The analysis of the effects on the structure and performance of lattice with different language models is given. Experiments show that the modified Katz approach enhances the system performance compared with traditional Katz approach. The best recognition rate achieves 60.90% for the corpus from interview program.
[1] Zheng Tieran, Han Jiqing. Syllable Lattice Based Chinese Speech Retrieval Techniques and Removing Redundancy Method from Indices. Acta Acustica, 2008, 33(6): 526-533 (in Chinese) (郑铁然,韩纪庆.基于音节Lattice的汉语语音检索技术及其索引去冗余方法.声学学报, 2008, 33(6): 526-533) [2] Zheng Tieran, Han Jiqing. Study on Chinese Speech Retrieval Based on Posterior Probability. Chinese High Technology Letters, 2009, 19(2): 119-124 (in Chinese) (郑铁然,韩纪庆.基于后验概率的汉语语音检索方法研究.高技术通讯, 2009, 19(2): 119-124) [3] Good I J. The Population Frequencies of Species and the Estimation of Population Parameters. Biometrika, 1953, 40(3/4): 237-264 [4] Katz S M. Estimation of Probabilities from Sparse Data for the Language Model Component of a Speech Recognizer. IEEE Trans on Acoustics, Speech, and Signal Processing, 1987, 35(3): 400-401 [5] Gale W A, Sampson G. Good-Turing Frequency Estimation without Tears. Quantitative Linguistics, 1995, 2: 217-237 [6] Yang Lin, Zhang Jianping, Yan Yonghong. Comparative Study on Smoothing Algorithms for Domain-Specific Chinese Language Models. Computer Engineering and Applications, 2006, 32(14): 14-16 (in Chinese) (杨 琳,张建平,颜永红.特定领域的汉语语言模型平滑算法比较研究.计算机工程与应用, 2006, 32(14): 14-16) [7] Zheng Tieran, Han Jiqing, Li Haiyang. Study on Performance Optimization for Chinese Speech Retrieval. Journal on Communications, 2009, 30(3): 84-88 (in Chinese) (郑铁然,韩纪庆,李海洋.基于词片的语言模型及在汉语语音检索中的应用.通信学报, 2009, 30(3): 84-88)
2011, Vol. 24 
