图书馆移动信息资源面向用户推荐优化仿真

摘要

An optimization method for user-oriented recommendation of library mobile information resource is proposed based on the partition clustering.Firstly,the rough set neighborhood theory is applied to the personalization recommendation of library mobile information resource,and the grading item set of user information resource is acquired.Then,the distances between the target user and all users in the matrix of user-information resource are worked out,and the nearest neighbor set of the target user is judged.Moreover,the dissimilarity matrix is introduced into the nearest neighbor calculation,and the user score of information resource in the nearest neighbor set is obtained.The library user set is also obtained integrated with the cluster partition theory,and the cluster between each library user attribute and core attribute is calculated.Finally,the grouping of approximate user is completed according to the user attribute,and the optimization for the user-oriented recommendation of library mobile information resource is achieved.The simulation results show that the proposed method improves the user-oriented recommendation precision and has good recommendation efficiency.%对图书馆移动信息资源面向用户推荐进行优化,可以有效解决随着图书馆移动信息资源剧增的问题.图书馆移动信息资源面向用户推荐时,应建立在获取用户对最近邻集合中信息资源项目的评分,在此基础上计算各个图书馆用户属性与核心属性之间的聚类的基础上,而传统方法利用图书馆书籍所属索引类别组建图书馆索引分布树进行推荐,到那时不能对图书馆用户属性与核心属性进行准确的聚类,导致信息资源面向用户推荐时耗时长,准确性差.提出一种划分聚类的图书馆移动信息资源面向用户推荐优化方法.上述方法先将粗糙集邻域理论应用到图书馆移动信息资源个性化推荐中,获取用户信息资源评分项集合,计算出目标用户与用户-信息资源项矩阵中所有用户对象之间的距离,由此判定目标用户的最近邻集合,将相异度矩阵概念引入到用户近邻计算中,给出用户对象的最近邻因子,得到用户对最近邻集合中信息资源项目的评分,结合聚类划分理论获取图书馆用户集合,并计算各个图书馆用户属性与核心属性之间的聚类,依据图书馆用户的属性对比完成近似用户的分组,由此完成图书馆移动信息资源面向用户推荐优化.仿真结果表明,所提方法可以有效提高图书馆移动信息资源面向用户推荐精度,具有较好的推荐效率.