Multi-behavior recommendation models excel in extracting abundant information from user-item interactions to enhance performance; however, they encounter challenges in accuracy due to noise disturbance and ambiguous weight allocation. In this paper, we propose cd-MBRec, a novel model designed to amplify commonality among various behaviors, thereby minimizing noise interference while preserving behavior diversity to highlight semantic variations in feedback across distinct scenarios. Specifically, the model begins by constructing behavior matrices that models separate behaviors, along with an interaction matrix offering a broad overview of user behaviors. It employs graph neural networks to extract higher-order semantic and structural information from input data. Concurrently, the model integrates principles of Weber-Fechner Law for the adaptive allocation of initial weights to the multiple behaviors and utilizes matrix factorization techniques for efficient behavior embedding. Extensive experiments on two real-world datasets demonstrate that cd-MBRec surpasses existing state-of-the-art models in recommendation performance, achieving notable average improvements of 4.96% in HR@10 and 7.75% in NDCG@10.

Multi-behavior recommendation has gained significant attention in recent years for its ability to outperform singlebehavior models.Current research related to multi-behavior models leaves room for improvement in the following two areas.First, the noise carried by individual behaviors and the additional noise generated during behavior processing is often overlooked, and these can ultimately degrade recommendation performance.Second, the specific time period of behavioral interactions and the frequency of interactions within that time period are also not taken into account.To address the above limitations, we propose a multibehavior recommendation model integrating dynamic preferences (MB-DP) that captures dynamic interests while smoothing and denoising multi-behavior information.MB-DP extracts low and high-order semantics from various behaviors and unifies the measurements to generate interaction predictions.Additionally, it analyzes the interaction time and frequency of each behavior using gated recurrent units to capture the dynamic preferences of users and improve the prediction values.Extensive experimental results on two real-world datasets show that MB-DP significantly improves recommendation performance compared to the state-ofthe-art baselines.

Multi-behavior recommendation has gained significant attention in recent years for its ability to outperform single-behavior models. Current research related to multi-behavior models leaves room for improvement in the following two areas. First, the noise carried by individual behaviors and the additional noise generated during behavior processing is often overlooked, and these can ultimately degrade recommendation performance. Second, the specific time period of behavioral interactions and the frequency of interactions within that time period are also not taken into account. To address the above limitations, we propose a multi-behavior recommendation model integrating dynamic preferences (MB-DP) that captures dynamic interests while smoothing and denoising multi-behavior information. MB-DP extracts low and high-order semantics from various behaviors and unifies the measurements to generate interaction predictions. Additionally, it analyzes the interaction time and frequency of each behavior using gated recurrent units to capture the dynamic preferences of users and improve the prediction values. Extensive experimental results on two real-world datasets show that MB-DP significantly improves recommendation performance compared to the state-of-the-art baselines.