摘要
为了更深入地挖掘食品安全事件中的特征和规律,本文提出了食品安全事件时间演化序列自动生成方法。首先从食品安全事件的时间构成、时间表达式左右边界特征、时间常用词三方面分析了食品安全事件时间的表达特征。基于时间表达式的多种特征,采用基于条件随机场模型的方法对时间表达式进行识别,F值最高达90%。然后以含有时间表达式的"句"为单位,将一个事件分为多个"时间-事件"的组合,最后采用基于时间戳的逆向匹配法生成事件时间演化序列。该方法可用于分析食品安全事件的发展脉络与特征,以期为政府实施食品安全事件监控提供理论与技术支撑。
A method for automatically generating time-evolution series food safety events is put forward to mine the characteristics and rules of food safety incidents in an in-depth manner. Then, expressing the time characteristics of a food safety event time is analyzed based on the time constitution, boundary characteristics, and common words used in food safety events. The time expression is identified using a conditional random field based on a variety of char- acteristics of time expressions. The F value was shown to reach 90%. The events are divided into a combination of multiple ＂time-incident＂ events based on the unit of sentences that contain the time expression. The time-evolution sequences are generated using a reverse matching method based on a timestamp. This method can be used to analyze the development and characteristics of a food safety incident and thereby provide theoretical and technical support for governments monitoring food safety events.
作者
王东波
叶文豪
吴毅
刘伙玉
苏新宁
沈思
Wang Dongbo Ye Wenhao Wu Yi Liu Huoyu Su Xinning Shen Si(College of lnformation and Technology, Nanjing Agricultural University, Nanjing 210095 School of Information Management, Nanjing University, Nanjing 210023 Jiangsu Key Laboratory of Data Engineering and Knowledge Service, Nanjing 210023 School of Economics ＆ Management, Nanjing University of Science and Technology, Nanjing 210094)
出处
《情报学报》
CSSCI
CSCD
北大核心
2017年第9期930-939,共10页
Journal of the China Society for Scientific and Technical Information
基金
江苏省社会科学基金"时间感知大数据特征下的食品安全突发事件应对策略挖掘研究"(15TQC003)
国家自然科学基金管理学部青年项目"基于CSSCI的句法级汉英平行语料库构建及知识挖掘研究"(71303120)
2011协同中心项目"面向应急推演平台的海量突发事件知识库与模型库构建研究"(JD20150101)
关键词
食品安全
时间演化
时间序列
条件随机场
时间戳
food safety
time evolution
time series
conditional random field
timestamp
