一种融合依存句法分析和语义角色标注的领域新闻热点话题识别方法  

A Method for Identifying Hot Topics in Domain News Texts Incorporating Dependent Syntax and Semantic Roles

作　　者：姜凌昊 王兴芬[1,2] JIANG Linghao;WANG Xingfen(School of Information Management,Beijing University of Information Technology,Beijing 100192,China;Business Big Data Analytics Research Center,Beijing University of Information Technology,Beijing 100089,China)

机构地区：[1]北京信息科技大学信息管理学院,北京100192 [2]北京信息科技大学商务大数据分析研究中心,北京100089

出　　处：《信息与电脑》2021年第19期64-67,共4页Information & Computer

基　　金：国家重点研发计划课题“大宗商品电子商务市场的交易风险智能分析与预警技术”(项目编号:2019YFB1405003)。

摘　　要：本文针对当前话题识别方法在发现热点话题中存在低效率、实时性不高等问题,提出一种融合依存句法分析和语义角色标注的领域新闻文本热点话题识别方法。实验数据集选用人工标注的领域新闻文本,将依存句法分析和语义角色标注用于主谓宾实体三元组抽取。对该方法分别和单纯的基于语义角色标注实体关系抽取以及MA-LDA主题抽取模型进行了对比实验,发现该方法的精确率、召回率、F值3个指标更优,尤其是含有动宾关系的领域新闻文本中,精确率达到90.64%。In this paper, we propose a method to identify hot topics in domain news text by integrating dependent syntactic analysis and semantic role annotation to address the problems of inefficiency and low real-time performance of current topic identification methods in discovering hot topics. The experimental dataset is manually annotated domain news text, and dependent syntactic analysis and semantic role annotation are used for subject-verb-object entity triad extraction. The accuracy, recall and F-value of this method are better than those of the pure semantic role annotation-based entity relationship extraction and MA-LDA topic extraction models, especially for the domain news text containing verb-object relationship, the accuracy rate reaches 90.64%.

关 键 词：话题检测与跟踪 依存句法 语义角色 TextRank 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]