基于对象阈值变精度概念格的电影推荐方法  

Recommended Method for Films Based on the Concept Grid of Object Threshold Variable Precision

作　　者：张国锋 杨凯[1] ZHANG Guofeng;YANG Kai(School of Computer and Software Engineering,Liaoning University of Science and Technology,Anshan 114051)

机构地区：[1]辽宁科技大学计算机与软件工程学院,鞍山114051

出　　处：《计算机与数字工程》2025年第3期671-677,746,共8页Computer & Digital Engineering

基　　金：国家自然科学基金项目(编号:51874171);辽宁省教育厅高等学校基本科研项目(编号:LJKZ0285)资助。

摘　　要：近些年,形式概念分析已经被应用于电影推荐系统。论文为解决其推荐精度过低及时间复杂度过大的问题,对形式概念分析的核心数据结构“概念格”进行改进。首先,利用评分信息将经典形式背景转化为模糊形式背景;然后,针对利用此模糊形式背景构造的Crisp-crisp变精度概念格进行改造,将作用于全局的变精度阈值k改造为阈值集合K,集合K中的每一个元素对应一位用户,作为此用户单独的变精度阈值;最后,便生成了新的Crisp-crisp变精度概念格结构,将其命名为对象阈值变精度概念格。在经典电影数据集上的一系列实验说明了这种结构的有效性。另外,针对的数据集中的电影类型数据进行处理,结合数学中悲观集与乐观集的性质,将电影类型数据转化为电影类型形式背景,生成电影类型概念格,最终结合对象阈值概念格与电影类型概念格共同计算相似度对用户进行推荐。实验证明,这种推荐方法要优于单独使用对象阈值变精度概念格进行推荐。In recent years,formal concept analysis has been applied to the film recommendation system,in order to solve the problem of its recommendation accuracy is too low and the time complexity is too large,the core data structure of formal concept analysis"concept grid"is improved.First,the use of scoring information to transform the classical form background into a blurred form background.Then,for the Crisp-crisp variable-precision concept grid constructed by using this fuzzy form background,the variable-precision threshold k acting on the global is transformed into a threshold set K,and each element in the set K corresponds to a user as a separate variable-precision threshold for this user.Finally,a new Crisp-crisp variable-precision conceptual lattice structure is generated,which is named the object threshold variable-precision conceptual lattice.A series of experiments on a classical film dataset illustrate the validity of this structure.In addition,the film type data in the data set is processed,combined with the nature of the pessimistic set and the optimistic set in mathematics,the film type data is converted into the film type form background,the film type concept grid is generated,and finally the similarity is calculated together with the object threshold variable precision concept grid and the film type concept grid to recommend the user.Experiments have proved that this recommendation method is superior to using the object threshold variable precision concept grid alone for recommendation.

关 键 词：形式概念分析 推荐系统 对象阈值变精度概念格 电影类型概念格 电影推荐 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]