网络答疑本体的生成与匹配方法研究

Study on generation and match of question answering ontology

针对网络答疑具体任务,提出答疑本体概念,构建答疑本体,并提出一种基于答疑本体的答案匹配方法和相似度计算公式。公式不仅考虑词之间的相似度,还考虑问题答案对(QAp)在本体中的位置。该位置信息隐含了句子结构上的语义和词的语义扩展。实验结果表明提出的答疑本体能够有效表示QAp,易于答疑系统的检索匹配。

Aiming at the requirements and traits of question answering systems, designed a structure of question answering ontology, which was the representation of natural language questions, and provided a well retrieval mechanism. Meanwhile, proposed an approach of ontology acquisition, which could decrease the costs of development manually. Finally, proposed formula in order to computer semantic similarity between sentences basing on question answering ontology. Now, the proposed ontology has already been implemented and applied in the natural language oriented Web answer system （ NL-WAS）, it shows that the question answering ontology can improve the result of the question search of NL-WAS system effectively.