关键词: 词汇-语义模式, 信息抽取, 金融事件, 词向量, 词列表, 概念词典

Abstract: Information Extraction is one of the most important tasks in NLP, in order to solve the problems of Information Extraction which caused by the diversity, ambiguity and structure of natural language, this paper proposed an approach of financial event information extraction based on hierarchical lexical-semantic patterns. In this method, a financial event model is defined at first, and secondly we employ a deep learning based method –word vector to obtain concept gazetteer automatically, in the last some hierarchical lexical-semantic patterns which drivened by finite state machine are used to extract some factors of financial events. The experiment shows that this method can extract the event information flexibly and effectively; the recognition precision of 26 types of financial events reached to 95.9%, F1 value reached to 91%.

Key words: Lexical-Semantic Patterns, Information Extraction, Financial Event, Word Vector, Word List, Gazetteer