问答系统命名实体识别改进方法研究

问答系统是一种以准确且自然的语言来回答用户提问的系统。本文对其中涉及的“命名实体识别”这一环节尝试了一些改进措施：1.针对传统单向模板匹配耗时耗力的问题,提出一种双向格子结构的长短时记忆网络（Lattice Bi-LSTM）,解决了命名实体识别中对句子处理不当和对分词结果具有依赖性两大问题,且与单向结构相比,双向结构能更好地利用句子信息,使输出结果更具鲁棒性,从而更准确地捕获语义信息。2.针对传统方法未考虑实体间相似度的非线性耦合性问题,提出一种利用周期性核函数准确地将“相似”的实体链接到知识库中去的方法。对提出的两个改进方法进行了实验验证,其结果表明：所用方法与经典方法相比,具有显著改进效果。

Research on the Improved Method of Named Entity Recognition in Q & A System

Q & A system is a kind of system which can answer user''s questions with accurate and natural language. Some improvement measures have been tried for "named entity recognition": 1. Aiming at the time-consuming and labor-consuming problem of traditional one-way template matching, this paper proposes a bidirectional lattice structure of short and long-term memory network (lattice Bi-LSTM), which solves the problems of improper sentence processing and dependence on the result of word segmentation in named entity recognition. Compared with unidirectional structure, bi-directional structure can make better use of sentence information and make the output more robust, thus capturing semantic information more accurately. 2. To solve the problem of non-linear coupling of similarity between entities in traditional methods, a method is proposed to link "similar" entities to the knowledge base accurately by using periodic kernel function. The two improved methods are verified by experiments, and the results show that the improved methods have significant improvement effects compared with the classical method.