A Survey on Statistical Language Models
统计语言模型综述

The Statistical Language Model (SLM) is a distribution to capture the generation rule in natural language. Since the first model was proposed in 1980, the SLM was used in many application such as speech recognition, optic character recognition, machine translation and etc. Recently the information retrieval based-on SLM is becoming a new direction in IR community. In this paper, we introduce the primary theory and current method about the SLM, and point out several promising research direction.