A Research on Sentence Similarity for Question Answering System Based on Multi-feature Fusion

Haipeng Ruan, Yuan Li, Qinling Wang, Yu Liu

科研成果: 书/报告/会议事项章节会议稿件同行评审

11 引用 (Scopus)

摘要

If just consider one feature of sentences to calculate sentences similarity, the performance of system is difficult to reach a satisfactory level. This paper presents a method of combining the features of semantic and structural to compute sentences similarity. It first discusses the methods of calculating the semantic similarity of sentences through word embedding and Tongyici Cilin. Next, it discusses the methods of calculating the morphological similarity and order similarity of sentences, and then combines the features through the neutral network to calculate the total similarity of the sentences. We include results from an evaluation of the system's performance and show that a combination of the features works better than any single approach.

源语言英语
主期刊名Proceedings - 2016 IEEE/WIC/ACM International Conference on Web Intelligence, WI 2016
出版商Institute of Electrical and Electronics Engineers Inc.
507-510
页数4
ISBN(电子版)9781509044702
DOI
出版状态已出版 - 12 1月 2017
活动2016 IEEE/WIC/ACM International Conference on Web Intelligence, WI 2016 - Omaha, 美国
期限: 13 10月 201616 10月 2016

出版系列

姓名Proceedings - 2016 IEEE/WIC/ACM International Conference on Web Intelligence, WI 2016

会议

会议2016 IEEE/WIC/ACM International Conference on Web Intelligence, WI 2016
国家/地区美国
Omaha
时期13/10/1616/10/16

指纹

探究 'A Research on Sentence Similarity for Question Answering System Based on Multi-feature Fusion' 的科研主题。它们共同构成独一无二的指纹。

引用此