A Chinese word dividing algorithm based on statistical language models

In Chinese speech processing, word dividing still remains a challenging problem, but it is one of the key problems for Chinese speech understanding, text-to-speech (TTS) systems and vocoders based on speech recognition and synthesis (SRSB vocoders). This paper proposes a word dividing algorithm based on statistical language models and Markov chain theory, which find the optimal word dividing solution without making use of grammatical and semantic knowledge.