基于SVM的网络信息茧房层次敏感影响因素识别研究  

Research on SVM-based Network Information Cocoons Hierarchy Sensitive Influencing Factors Identification

作　　者：王益成[1] 王萍[1] 王美月 Wang Yicheng;Wang Ping;Wang Meiyue(Management School of Jilin University,Changchun,130022)

机构地区：[1]吉林大学管理学院

出　　处：《情报资料工作》2019年第6期90-97,共8页Information and Documentation Services

基　　金：国家自然科学基金项目“政府网站信息资源多维语义知识融合研究”（编号：71740015）的研究成果

摘　　要：文章利用信息增益理论解析各影响因素与信息茧房层次之间的关联度,构建信息茧房层次敏感影响因素模型,进而通过支持向量机理论构建信息茧房层次预测模型.信息茧房层次的12个敏感影响因素集中于算法推荐技术、用户信息素养、系统交互性与用户行为特征四个子维度,模型分类预测精度为84.92%,表示支持向量机对信息茧房层次有良好的预测能力.The paper uses information gain theory to analyze the correlation between each influencing factor and thelevel of the information cocoons,builds the model of the sensitive influence factor on the level of the information co-coons,and then builds the prediction model of the level of the information cocoons through the support vector machinetheory.The 12 sensitive influencing factors of the level of information cocoons are concentrated in four sub-dimen-sions,including algorithm recommendation technology,user information literacy,and system interactivity and user be-havior characteristics.The accuracy of model classification prediction is 84.92%,indicating that the support vector ma-chine has good predictive ability for the level of information cocoons.

关 键 词：信息茧房 信息增益理论 支持向量机 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]