Research for Fuzzy Matching Algorithm of Key Semantic Class in Chinese Spoken Language Understanding
LI Yan-lin · Journal of Chinese Computer Systems · 2014
This paper focuses on obtaining key semantic concepts by fuzzy matching method to solve some problems such as words' variation induced by automatic speech recognition( ASR),or missing field of key information by users in the process of voice search.Two-stage method is proposed: first,conditional random field( CRF) is used for building probabilistic models to segment and label entity names from an input sentence. Second,similarity measures are computed on Chinese pinyin of both the named entities labeled by CRF model and reference characters of dictionary. Then a key semantic concept with maximum similarity will replace the wrong named entity. Furthermore,considering one-best fuzzy matching result may be not the one which users want to,experiments about10-best are conducted to provide more candidate results for people. We compared fuzzy matching algorithms based on different similarity functions. The experiment shows that similarity measures based on Chinese pinyin method can achieve higher performance than that of based on Chinese characters.