用k-最近邻和贝叶斯分类预测图书用户喜好  

Predicting users' interest of digital books using k-closest Neighbor and Naive Bayesian classification

作　　者：钟亮[1,2] 

机构地区：[1]广西大学计算机与电子信息学院,南宁530004 [2]广西大学行健文理学院,南宁530005

出　　处：《信息技术》2016年第9期62-65,共4页Information Technology

基　　金：广西科学研究与技术开发计划项目(桂科能1140008-3B)

摘　　要：将用户兴趣爱好、书籍受欢迎程度和用户对书籍的评价等参数组成差异性矩阵,确定各参数影响权重、建模计算用户对书籍的评分,融合采用k-最近邻分类法和朴素贝叶斯分类法来分类过滤数据,设计实现一种数字图书用户喜好预测算法。实验结果表明,该算法提高了数字图书个性化推荐精度。User's interests,the popularity of books and user evaluation parameters are used to generate the difference matrix of the books,the weights of the parameters are determined,users' scores of books are computed, and data are filtered by using k-nearest neighbor classification and Naive Bayes classification, and a predicting users ' interests of digital books algorithm is implemented. The experimental results show that the algorithm can improve the accuracy of the digital book personalized recommendation.

关 键 词：数字图书 个性化推荐 k-最邻近分类 朴素贝叶斯分类 内容过滤 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]