Extracting entities for natural language dialog system

Natural language dialog systems have achieved great success on certain tasks. But they can´t be ported to new domains easily. This paper talks about entity extraction, and presents some observations relevant to entity extraction, as well as the details of implementation of an entity extraction algorithm for Chinese text. This algorithm is divided into five steps. Parts of the algorithm include rules analysis, statistic analysis and syntax analysis. A primitive experiment is given, and proves that entity extraction is viable