摘要
将信息筛选描述为信息对象到用户偏好值的映射函数 ,利用多目标决策方法提出了一个信息筛选中群体用户偏好聚合模型 .对信息筛选过程进行了分析 ,它包括定义用户偏好、接受信息输入流、计算其用户偏好值以及用户偏好修正 .在此基础上 ,用偏好向量定义个体用户偏好 ,并通过实例进行演示 .系统研究了信息筛选中的群体用户偏好 ,建立了一个群体偏好聚合模型 ,计算了不考虑信息代价和考虑信息代价两种情况下的群体偏好值 .利用模型通过聚合个体偏好求取群体偏好 .在考虑信息代价时 ,对不同类别的信息代价进行归一化处理 .实例验证表明 。
By using multiple criteria decision making theory an aggregation model for group user profile in information filtering was proposed. The information filtering is defined as a function from information to user preference in which user profile reflects the regular interest, task and demand of the longtime users. The process of information filtering was analyzed. It includes defining user profile, accepting information inputting stream, calculating the value of use preference and adjusting user profile. Based on it, individual user profile was defined using profile vector. An example was given to demonstrate it. Furthermore, group profile was studied systematically. An aggregation model was proposed for group profile both with and without consideration of information charge. Using this model, group profile was got by aggregating individual profiles. Under consideration of information charge, different types of charge were normalized. The model was explained in details by several examples.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2000年第6期818-820,共3页
Journal of Shanghai Jiaotong University
基金
上海交通大学基金资助项目
