摘要
为应对推荐系统中数据日益增长的现状,在均模型的基础上,提出支持增量扩展的均模型算法(incremental mean model,Incremental MM)。均模型能够提高固定数据集的推荐效率,在均模型转换之前,为各项目建立评分值与评分数的评分映射表,通过映射表的更新实现均模型的增量算法。在MovieLens数据集上进行实验,实验结果表明,该算法的增量更新效率较高,无损推荐精度,指出了其在大数据背景下的性能优势,展现了其可观的应用前景。
In response to the growing status of data in the field of recommendation system,incremental mean model which supported incremental expansion was proposed on the basis of the mean model.The mean model improved the efficiency of the recommendation system well in fixed data sets.The mapping of score and number of the items was set up before mean model was built,and incremental mean model was implemented by updating the mapping table quickly.Results of experiments on MovieLens data sets show the proposed method has higher incremental update efficiency and no loss of recommendation accuracy.The analysis of the experimental results shows the advantages of the proposed algorithm in the background of large data,and demonstrates the considerable application prospect clearly.
出处
《计算机工程与设计》
北大核心
2017年第3期659-663,676,共6页
Computer Engineering and Design
基金
国家863高技术研究发展计划基金项目(2014AA015204)
关键词
均模型
增量扩展
大数据
映射
推荐系统
协同过滤
mean model
incremental extension
big data
map
recommender system
collaborative filtering
