[目的/意义]针对目前基于网络的话题识别与分析方法的局限性,提出针对网络问答社区的话题识别与分析方法,为此类网站的话题识别与分析提供参考。[方法/过程]以改进的中文分词技术为基础,构建网络问答社区的话题识别指标,通过线性加权方式计算权重,结合关键词提取方法确定话题关键词,对话题关注焦点进行提取,对分布情况进行测度。依据所提出的改进方法,以知乎网站为数据来源,从话题关键词、关键词分布以及热点子话题3个角度对“老年人”话题焦点进行识别与分析。[结果/结论]研究表明,该方法具有科学性和可行性,不仅拓展了社会问题的分析数据源,也为“积极开展应对人口老龄化行动”提供了决策依据。
[Purpose/significance]This paper aims at the limitations of current methods and proposing a topic detection and analysis method for the social Q&A website, which provides a reference for this website in topic detection and analysis. [ Method/process] This method is based on the improved Chinese word segmentation techniques, using the Linear Weighted to determine the weight of words, combined with keyword extraction method for determining topic keywords to extract the topic fo- cus and measure the distributions. According to the proposed method, based on data from Zhihu, we carry out the network topic detection and analysis of the topic "the elderly" from three angles: the focus of topic, topic distribution and hot subtopics. [ Resuit/conclusion]This study has shown that this method is scientific and practical. It extends the analytical data source of social problems and provides a basis of decision- makingfor the "Actively Deal with the Population Ageing" activity.