基于对象阈值变精度概念的启发式推荐方法
Heuristic recommendation method based on concept of variable precision of object thresholds
张国锋 1杨凯1
作者信息
- 1. 辽宁科技大学计算机与软件工程学院,辽宁鞍山 114051
- 折叠
摘要
针对启发式概念推荐方法进行优化,提出一种基于对象阈值变精度概念的启发式推荐方法.将模糊形式背景转化为对象阈值可变的变精度形式背景,在此形式背景的基础上定义概念构造的启发式信息;利用内涵约束,在保证群组相似度的基础上,构造当前加权面积最大的概念,将这些概念组成强概念集合;在这个概念集合中利用组推荐方法对所有用户进行推荐.实验在MovieLens系列数据集和FilmTrust数据集上的一系列结果验证这种优化是有效的,能够大幅加强推荐的精度和性能.
Abstract
Optimized for the heuristic concept recommendation method,a heuristic recommendation method based on the concept of variable accuracy of object threshold was proposed.The blurred formal background was transformed into a variable-precision formal background with variable object thresholds,and the heuristic information of conceptual construction was defined on the basis of this formal background.By using the connotation constraint,on the basis of ensuring the similarity of the group,the concepts with the largest weighted area were constructed,and these concepts were formed into a strong concept set.The group recommendation method was used in this conceptual collection to recommend all users.A series of experimental results on the MovieLens series dataset and the FilmTrust dataset verify that this optimization is effective,greatly improving the accuracy and performance of recommendations.
关键词
形式概念分析/概念格/变精度概念/对象阈值变精度概念/强概念/启发式算法/推荐系统Key words
formal concept analysis/conceptual lattice/variable precision concept/object threshold variable precision concept/strong concept/heuristic algorithms/referral system引用本文复制引用
基金项目
国家自然科学基金项目(51874171)
辽宁省教育厅高等学校基本科研基金项目(LJKZ0285)
出版年
2024
NETL
NSTL
万方数据