因素空间中的多标准因果分析方法  

Multi-target causality analysis in factor space

作　　者：曲卫华 刘海涛 郭嗣琮 

机构地区：[1]太原学院计算中心,山西太原030032 [2]辽宁工程技术大学智能工程与数学研究院,辽宁阜新123000

出　　处：《模糊系统与数学》2017年第6期74-81,共8页Fuzzy Systems and Mathematics

基　　金：国家自然科学基金资助项目(61350003;11401284);教育部高校博士学科点专项科研基金资助项目(20102121110002);辽宁省教育厅科学技术研究一般基金资助项目(L2014133);山西省高等学校哲学社会科学研究项目(20172206)

摘　　要：为了进一步扩大因素空间中的因果分析法的应用领域,本文提出了多标准因果分析算法。因果分析是从多个条件因素到一个结果因素的分析,多标准因果分析则研究从多个条件因素到多个结果因素的因果分析。从应用领域的拓展来说,这一方法的应用可从多属性单体决策拓展到多属性群体决策,从单向归纳学习拓展到多向归纳学习,从单标签分类拓展到多标签分类.本文所建立的算法简捷,适于大数据的应用。In order to extend the applied areas of causality analysis in advance, the paper presents multi-standard causality analysis method. There was one standard factor resulted from a group of conditional factors by causality analysis, but there is a group of multi-standard factors resulted from a group of conditional factors by multi-standard causality analysis. From the view of extension of applied areas, the new method extends areas from one person＇s multi-attribute decision making to a group of makers＇ related decision making, from one dimension of induction/learning to high dimension of induction/learning, from unique label classification to multi-label classification. Relative algorithm is simple and fast for the implication facing big data.

关 键 词：因素空间 因果分析 多标签分类 

分 类 号：C931.1[经济管理—管理学]