User-Specific Feature-Based Similarity Models for Top- n Recommendation of New Items

Author:

Elbadrawy Asmaa1,Karypis George1

Affiliation:

1. University of Minnesota, Minneapolis, MN

Abstract

Recommending new items for suitable users is an important yet challenging problem due to the lack of preference history for the new items. Noncollaborative user modeling techniques that rely on the item features can be used to recommend new items. However, they only use the past preferences of each user to provide recommendations for that user. They do not utilize information from the past preferences of other users, which can potentially be ignoring useful information. More recent factor models transfer knowledge across users using their preference information in order to provide more accurate recommendations. These methods learn a low-rank approximation for the preference matrix, which can lead to loss of information. Moreover, they might not be able to learn useful patterns given very sparse datasets. In this work, we present <scp>UFSM</scp>, a method for top-<i>n</i> recommendation of new items given binary user preferences. <scp>UFSM</scp> learns <b>U</b>ser-specific <b>F</b>eature-based item-<b>S</b>imilarity <b>M</b>odels, and its strength lies in combining two points: (1) exploiting preference information across all users to learn multiple global item similarity functions and (2) learning user-specific weights that determine the contribution of each global similarity function in generating recommendations for each user. <scp>UFSM</scp> can be considered as a sparse high-dimensional factor model where the previous preferences of each user are incorporated within his or her latent representation. This way, <scp>UFSM</scp> combines the merits of item similarity models that capture local relations among items and factor models that learn global preference patterns. A comprehensive set of experiments was conduced to compare <scp>UFSM</scp> against state-of-the-art collaborative factor models and noncollaborative user modeling techniques. Results show that <scp>UFSM</scp> outperforms other techniques in terms of recommendation quality. <scp>UFSM</scp> manages to yield better recommendations even with very sparse datasets. Results also show that <scp>UFSM</scp> can efficiently handle high-dimensional as well as low-dimensional item feature spaces.

Funder

Digital Technology Center at the University of Minnesota

NSF

Publisher

Association for Computing Machinery (ACM)

Subject

Artificial Intelligence,Theoretical Computer Science

Cited by 41 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

1. TRANSGNN: Towards Knowledge Enhanced Top-K Recommendation via Transformed Heterogeneous Graph Neural Network;2023 IEEE International Conference on Big Data (BigData);2023-12-15

2. IGCN: Item Influence Enhanced Graph Convolution Networks for Recommendation of Cold-Start Items;2023 IEEE International Conference on Data Mining Workshops (ICDMW);2023-12-04

3. FGCR: Fused graph context-aware recommender system;Knowledge-Based Systems;2023-10

4. An Integrated Support System for People with Intellectual Disability;Electronics;2023-09-08

5. A Content Recommendation Platform for People with Intellectual Disability;2023 Fourteenth International Conference on Ubiquitous and Future Networks (ICUFN);2023-07-04

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3