项目名称: 基于聚合的社会化短文本信息处理与细粒度倾向性分析
项目编号: No.71503126
项目类型: 青年科学基金项目
立项/批准年度: 2016
项目学科: 管理科学
项目作者: 薛春香
作者单位: 南京理工大学
项目金额: 17万元
中文摘要: 面向基于社会化媒体信息的细粒度倾向性分析的实际需求,针对社会化短文本信息由于文字简短导致的语境缺失、特征稀疏问题,本研究提出基于多层次、多维度信息聚合解决方案。本项目综合运用实证分析、模型构建和原型系统验证等研究方法,梳理和描述社会化短文本信息属性特征,剖析从数据整合-内容聚合-结果呈现不同阶段社会化短文本信息处理和分析的聚合要求,探索基于内容聚合和社会化属性特征聚合的社会化短文本信息语境生成,构建基于主题相关性的倾向性分析和融合人群特征的群体倾向性分析模型。本项目采用多层次、多维度的聚合实现短文本信息的语境生成和情感语义分析,融合内容分析和社会化属性特征进行细粒度倾向性分析,为社会化短文本信息处理与细粒度倾向性分析提供新的思路和方法。
中文关键词: 社会化短文本;信息聚合;倾向性分析;智能信息处理;社会化媒体
英文摘要: Starting from the demand for fine-grained orientation analysis of online users based on social media, aiming at the problems of data sparseness and lack of contextual information of large amounts of short texts, this project proposes the new solution of multi-level and multi-dimensional information aggregation for short texts which were generated in the social media. Through empirical analysis, model construction and prototype system, the proposed project summaries and describes the attributes and features of social media information, analyzes the information aggregation needs at the different stages of data integration, content aggregation and result presentation, generates contexts for social short texts based on information aggregation of contents and attributes, and constructs topic-related and group orientation analysis models. This research provides a new idea and method for processing and orientation analysis of social media information based on multi-level and multi-dimensional aggregation with their contents and social attributes.
英文关键词: social short texts;information aggregation;orientation analysis;intelligent information processing;social media