Article
Computer Science, Information Systems
Zhang Li, Chen XiaoBo
Summary: The paper proposes a recommendation algorithm based on social networks, which optimizes the recommendation results by considering the trust relationships and social influence between users. The experiments show that the algorithm has significant advantages in recommendation accuracy.
MULTIMEDIA TOOLS AND APPLICATIONS
(2022)
Article
Computer Science, Hardware & Architecture
Shangshang Xu, Haiyan Zhuang, Fuzhen Sun, Shaoqing Wang, Tianhui Wu, Jiawei Dong
Summary: This paper proposes a hybrid method based on probabilistic matrix factorization and directed trust to improve the performance of recommender systems, addressing the sparsity of trust matrix and capturing trust relations among users. Experimental results demonstrate that the proposed algorithm outperforms existing benchmark algorithms.
COMPUTERS & ELECTRICAL ENGINEERING
(2021)
Article
Computer Science, Information Systems
Maryam Jallouli, Sonia Lajmi, Ikram Amous
Summary: This paper proposes a recommendation model called REMOVE based on socio-environmental context to address the new user and data sparsity problems in recommender systems. By integrating social and environmental information, the model aims to improve recommendation quality by modifying the characteristics of both items and users. Experimental results show that REMOVE outperforms other approaches in terms of Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE) and effectively handles the identified issues.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Jyoti Shokeen, Chhavi Rana
Summary: This paper proposes a trust and semantic-based social recommendation approach to address cold-start issues, utilizing social relationships to compute trust and extracting implicit data. By discovering top-k semantic friends in social networks, the approach outperforms traditional methods that give equal weights to all users.
JOURNAL OF AMBIENT INTELLIGENCE AND HUMANIZED COMPUTING
(2021)
Article
Computer Science, Artificial Intelligence
Yunfei He, Yiwen Zhang, Lianyong Qi, Dengcheng Yan, Qiang He
Summary: With the advancement of the internet, recommendation systems are utilizing complex data for improved performance. A new embedding method called HopRec is proposed to enhance recommendation by capturing relationships through outer product, showing significant superiority in experiments over existing methods.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Information Systems
Lijuan Weng, Qishan Zhang
Summary: With the rapid development of information technology, social media has become widely used, leading to information overload for consumers. This study introduces a new social recommendation method based on opinion leaders, demonstrating superior performance on real datasets.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Review
Computer Science, Information Systems
Rui Chen, Kangning Pang, Min Huang, Hui Liang, Shizheng Zhang, Lei Zhang, Pu Li, Zhengwei Xia, Jianwei Zhang, Xiangjie Kong
Summary: With the development of online social networks, more and more users are participating and forming rich social relationships. These relationships provide a data source and research basis for recommender systems, driving the development of recommender systems based on social networks.
Article
Computer Science, Artificial Intelligence
Xue Zhang, Bin Wu, Yangdong Ye
Summary: Recommender systems often face the challenges of data sparsity and cold start. Existing studies have limitations in model design and model learning. To address these limitations, we propose a new probabilistic method called GAMF, which captures high-order social relationships and uses an attention mechanism to differentiate influences between friends. Our method also utilizes an efficient optimization algorithm for parameter learning.
Article
Computer Science, Artificial Intelligence
Yuan-Yuan Xu, Shen-Ming Gu, Fan Min
Summary: This paper proposes an efficient and effective outlier removal algorithm to improve the quality of training data. By modeling the noise as a mixture of Gaussian distribution and calculating low-rank matrices, the algorithm compares the original and recovered ratings to identify suspected outliers.
INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS
(2022)
Article
Computer Science, Artificial Intelligence
Junrui Liu, Zhen Yang, Tong Li, Di Wu, Ruiyi Wang
Summary: This paper proposes a novel personalized recommendation method called similarity pairwise ranking (SPR) to address the issue of imbalanced data distribution affecting the effectiveness of Bayesian personalized ranking (BPR). By eliminating the score differences between popular and personalized items based on their similarity, SPR enhances the recommendation quality and better meets the individual needs of users. Experimental results demonstrate the superiority of SPR over recent state-of-the-art methods.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Computer Science, Information Systems
Shulin Cheng, Huimin Jiang, Wanyan Wang, Wei Jiang
Summary: Compared to traditional recommender systems, context-aware recommender systems are better suited to real-world application contexts. However, most existing research has focused on single context-aware recommendations, such as time or location, and lacks in-depth analysis of multi-context-aware recommendations. Therefore, we proposed a high-order tensor factorization recommendation method based on multi-context awareness. By detecting user sensitivity to multiple contexts and constructing four-dimensional tensors and feature matrices, we were able to effectively address data sparsity. Our method improved recommendation quality through parameter optimization and filling in missing data, and was validated using a multi-context-aware movie dataset.
MULTIMEDIA SYSTEMS
(2023)
Article
Engineering, Multidisciplinary
Liangtian Wan, Feng Xia, Xiangjie Kong, Ching-Hsien Hsu, Runhe Huang, Jianhua Ma
Summary: Recent years have seen a surge in information overload on online social networks, leading to increased interest in social network based recommender systems. This study introduces a novel trust-aware approach based on deep learning to improve recommendation performance, incorporating deep matrix factorization techniques, deep marginalized denoising autoencoder, and community regularization, which outperformed existing baselines, particularly for cold-start users.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2021)
Article
Computer Science, Artificial Intelligence
Joojo Walker, Fengli Zhang, Fan Zhou, Ting Zhong
Summary: Most existing studies in recommender systems assume socially connected users have equal influence, but this is not practical. The SOAP-VAE model captures varying levels of influence and interaction patterns among entities to effectively alleviate data sparsity. Empirical results show SOAP-VAE outperforms previous models on real-world datasets.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2022)
Article
Computer Science, Information Systems
Naina Yadav, Sukomal Pal, Anil Kumar Singh, Kartikey Singh
Summary: This paper introduces the collaborative filtering method for recommender systems, but points out the issues in diversity and coverage. To address this problem, the authors propose a cluster-based diversity recommendation method, which utilizes different clustering techniques and pre-trained models for generating diverse recommendations.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Article
Computer Science, Artificial Intelligence
Le Nguyen Hoai Nam
Summary: This paper presents two different latent factor models, SC1 and SC2, for integrating user implicit and explicit preferences. SC1 works well for experienced users while SC2 is suitable for users with less domain knowledge.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)