项目名称: 信息不完全的双边匹配决策方法研究
项目编号: No.71501023
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: 张震
作者单位: 大连理工大学
项目金额: 15.4万元
中文摘要: 现实生活中存在着大量的双边匹配决策问题,但已有研究考虑的多是匹配主体的评价/偏好信息完全给出的情况。匹配主体在给出评价/偏好信息时,通常会给出关于对方的不完全信息。本项目以已有双边匹配决策研究成果为基础,将数据挖掘中的缺失值填补算法和推荐系统中的协同过滤算法引入双边匹配决策领域,对信息不完全的双边匹配决策方法进行研究,主要研究内容包括:(1)基于不完全序值信息和得分的双边匹配决策方法;(2)信息不完全的多指标双边匹配决策方法;(3)基于不完全偏好关系的双边匹配决策方法;(4)基于不完全信息的双边匹配决策支持原型系统设计与开发。 . 本项目的研究具有重要的理论意义和实用价值,其研究成果不但可以丰富双边匹配决策理论与方法的研究,而且可以广泛应用于解决电子商务中的交易匹配和人力资源管理中的员工-岗位匹配等实际匹配决策问题。
中文关键词: 双边匹配;不完全信息;多属性决策;偏好关系;推荐系统
英文摘要: Two-sided matching decision making problems exist widely in real life. However, most of the previous research focuses on the situation in which matching objects' evaluation/preference information on the other side is completely given. When giving evaluation/preference information on matching objects of the other side, incomplete information is usually provided. On the basis of previous research results for two-sided matching decision making, this project intends to investigate two-sided matching decision making methods with incomplete information by combining missing values imputation methods in data mining and collaborative filtering algorithms in recommender systems. The main research work consists of four parts: (1) two-sided matching decision making methods based on incomplete preference ordinals and incomplete scores; (2) multi-attribute two-sided matching decision making methods with incomplete information; (3) two-sided matching decision making methods based on incomplete preference relations; (4) prototype system design and development for two-sided matching decision support with incomplete information.. The research is of great theoretical significance and practical values. The results of this project can not only enrich the research of two-sided matching decision making theory and methods, but also be used to deal with practical matching decision making problems, such as trade matching in e-commerce and employee-position matching in human resource management.
英文关键词: two-sided matching;incomplete information;multi-attribute decision making;preference relation;recommender system