统计与决策2024，Vol.40Issue(4)：23-27,5.DOI:10.13546/j.cnki.tjyjc.2024.04.004
一种修正评分偏差并精细聚类中心的协同过滤推荐算法
A Collaborative Filtering Recommender Algorithm for Correcting Rating Biases and Refining Cluster Centers
摘要
Abstract
As one of the widely studied recommender algorithms by scholars globally,collaborative filtering is adversely af-fected by issues such as rating biases and data sparseness,leading to suboptimal recommendation performance.In order to address the aforementioned issues,this paper proposes an improved clustering-based collaborative filtering recommender algorithm.First-ly,the algorithm utilizes unsupervised sentiment mining techniques to map the sentiment of comments into a value within a fixed range,correcting user rating biases through weighted adjustments.Then,the algorithm constructs a data field for the modified us-er-item rating matrix,employes a heuristic optimization algorithm to calculate the optimal number of clusters and the best initial clustering centers.This facilitates the clustering segmentation of users,integrating proximity-based user similarity and ratings to generate recommendations.Finally,the performance and effectiveness of the proposed algorithm are comprehensively evaluated based on three self-constructed real-world datasets.The experimental results show that the improved algorithm outperforms other algorithms in terms of Precision,Recall and F1-Score evaluation indexes,and proves effective in addressing data sparsity and im-proving the recommendation performance of the recommender system.关键词
评分偏差/随机初始聚类中心/协同过滤/评论情感挖掘/数据场聚类Key words
rating biases/random initial clustering centers/collaborative filtering/comment sentiment mining/data field clustering分类
信息技术与安全科学引用本文复制引用
马鑫,段刚龙..一种修正评分偏差并精细聚类中心的协同过滤推荐算法[J].统计与决策,2024,40(4):23-27,5.基金项目
陕西省软科学项目(2022KRM188) （2022KRM188）
