项目名称: 快速分析网络食品安全问题的歧义性的若干关键技术研究
项目编号: No.61303214
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 自动化技术、计算机技术
项目作者: 刘金硕
作者单位: 武汉大学
项目金额: 23万元
中文摘要: 网络舆情是网络内容安全的重要的研究内容,对网络舆情的监测与分析具有重要的理论研究意义。本系统以快速识别及析取食品领域质量安全信息特征词及其消歧方法的研究为内容,实现对食品质量安全信息的精确监测和分析,提出一个多级并行粒度的倾向性分析与食品安全消歧的快速分析的框架NPO-OA;通过灵活的多级并行粒度的并行方法,提高倾向性分析与食品安全问题消歧的效率;同时,设计基于文本切片的上下文句式模型训练,以实现专有名词快速识别和析取模型方法。以及基于特定评估函数实现的特征词消歧方法,提升网络舆情的获取的效率以及话题检测的准确度。本项目旨在探索基于网络舆情的食品安全分析问题时,快速明确食品安全问题,以及快速消除食品安全问题中的歧义性问题规律,为快速判断网络舆情信息与精确实现倾向性分析打下基础。
中文关键词: 歧义消除;食品安全;舆情分析;内容安全;信息安全
英文摘要: Online public opinion is a major part of content security in the Internet. The monitor and analysis of online public opinion is of great important significance. A major content of this research is recognizing feature words of security information in food industry, to achieve exact monitor and analysis of security information in food quality, by using a multilevel parallel frame named NPO-OA. Another major content is presenting a flexible multilevel parallel method to improve the efficiency of obtaining online public opinion and opinion analysis. Meanwhile, this research puts forward an optimized technology to discriminate proper noun based on syntactical structure of labled sentences of sliced context. We also present a disambiguation method for feature words based on specific evaluation function, to promote the efficiency of obtaining online public opinion and the accuracy rating of topic detection. This research aims at exploring the regular pattern of quickly determining the food security problem,using the internet opinion anlysisi techniques, and deleting the ambiguity of food security. Our research can set up the basement for quickly determining internet opionion, and accurately realize the trend anlysis.
英文关键词: ambiguity deduction;food safety;public opionion suoervision;contenct safety;information safety