Learning coupled latent features via review texts for IOT service recommendation. (July 2022)

• Record Type: Journal Article
• Authors: Zhang, Quangui; Wang, Li; Xu, Keda; Lu, Wenpeng; Ma, Xinqiang; Huang, Yi
• Is Part Of: Computers & electrical engineering. Volume 101(2022)
• Publication Date: 2022-07
• Subjects: Attention mechanism -- Coupling learning -- Convolutional neural network -- Recommender systems; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2022.108084
• Languages: English
• ISSNs: 0045-7906

Abstract

Abstract: The existing Internet of Things (IOT) service recommendation models are generally based on collaborative filtering algorithms. However, they face challenges in achieving better recommendations. The main reason is that the user-service interaction matrix is based on the assumption of independent and identical distribution, which ignores the characteristics of the users and services. To this end, we propose a coupled latent feature learning model learning the coupling relationship between users/services. The model learns user/service intra-couplings showing the relationships between user/service explicit and latent features by a convolutional neural network. It then learns user-service inter-couplings. This refers to the relationships between user and service features by an attentional multi-layer perceptron in which an attention layer captures varying feature attention vectors. Through extensive experiments on two real-world datasets, including Amazon Movies and TV and Yelp, experimental results demonstrate that the proposed model outperforms current state-of-the-art methods. More specifically, the proposed method outperforms other conventional methods by 30%. Graphical abstract: Highlights: CoupledLFL learns user/service intra-couplings by a convolutional neural network. CoupledLFL learns user-service inter-couplings by multi-layer networks. CoupledLFL essentially explains why a user likes an item.