一种联合LTR和社交网络的Top-k推荐方法  

Approach for Top-k Recommendation Based on LTR and Social Network

作　　者：熊丽荣[1] 王玲燕 黄玉柱[2] XIONG Li-rong;WANG Ling-yan;HUANG Yu-zhu(College of Computer Science and Technology,Zhejiang University of Technology,Hangzhou 310023,China;Zhejiang Sci-Tech University,Hangzhou 310018,China)

机构地区：[1]浙江工业大学计算机科学与技术学院,杭州310023 [2]浙江理工大学,杭州310018

出　　处：《小型微型计算机系统》2018年第12期2577-2584,共8页Journal of Chinese Computer Systems

摘　　要：个性化推荐可以有效地解决网络中数据爆炸问题,绝大多数的工作利用用户评分信息来降低评分预测误差值(MAE).相比于精确的评分预测,用户更关心推荐给他的top-k项目排序列表,而最小化MAE并不意味着良好的top-k项目列表.由于评分数据存在稀疏问题,社交网络中信任信息在top-k推荐系统中起着越来越重要的作用.本文提出了一种基于信任的面向top-k排序的推荐方法,BTRank.该算法基于LTR方法,结合用户评分以及用户信任信息来构建项目排序模型,有效地提高对所有用户的top-k排序列表质量.同时,考虑到用户兴趣会随着时间演变而变化,本文设计了时间效应模型函数用于处理用户历史评分数据.在真实数据集上的实验中表明本文提出的算法效果明显优于传统的推荐算法以及同类top-k排序推荐算法.Personalized recommendations can effectively solve the data explosion problem in network. Most existing works utilize rating information to reduce the score prediction error,e. g. MAE;however,users prefer sorted top-k item lists and minimizing MAE criteria does not always result in better top-k item lists. Meanwhile,social connections among users play an increasingly important role in top-k recommendation system to alleviate data sparse problem. So,this paper presents a trust-based top-k recommendation algorithm,called BTRank. Based on LTR method,it integrates rating and trust information to construct the rating sorting model,which effectively improves the quality of the top-k item list of all users. And considering the user interest will evolve over time,we design a time attenuation model to process users＇ item ratings and trust ratings. A series of experiments on real world datasets prove the effectiveness of our algorithm.

关 键 词：社会化推荐 学习排序 矩阵分解 时间衰减 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]