From Individual to Group: Developing a Context-Aware Multi-Criteria Group Recommender System

TL;DR

This work addresses the challenge of group decision-making by proposing a Context-Aware Multi-Criteria Group Recommender System (CA-MCGRS) that leverages a Multi-Head Attention (MHA) architecture to fuse group, item, context, and criteria features. It formalizes the problem with sets $U$, $G$, $I$, $C$, $CR$, and $R$, and optimizes top-$K$ group recommendations by maximizing the sum of context- and criteria-dependent ratings $r(g,i,c,cr)$. The model architecture includes embedding layers for inputs, a concatenation step, an MHA module, and dense prediction layers, with a regression objective minimized via Mean Squared Error and trained using Adagrad. Experiments on the ITM-Rec educational dataset show that CA-MCGRS with MHA outperforms several strong baselines across multiple scenarios, demonstrating the value of incorporating context and multi-criteria evaluations for group recommendations. The findings suggest significant implications for educational settings and other domains requiring nuanced group-consensus recommendations, with future work focusing on richer context integration, scalability, and cross-domain validation.

Abstract

Group decision-making is becoming increasingly common in areas such as education, dining, travel, and finance, where collaborative choices must balance diverse individual preferences. While conventional recommender systems are effective in personalization, they fall short in group settings due to their inability to manage conflicting preferences, contextual factors, and multiple evaluation criteria. This study presents the development of a Context-Aware Multi-Criteria Group Recommender System (CA-MCGRS) designed to address these challenges by integrating contextual factors and multiple criteria to enhance recommendation accuracy. By leveraging a Multi-Head Attention mechanism, our model dynamically weighs the importance of different features. Experiments conducted on an educational dataset with varied ratings and contextual variables demonstrate that CA-MCGRS consistently outperforms other approaches across four scenarios. Our findings underscore the importance of incorporating context and multi-criteria evaluations to improve group recommendations, offering valuable insights for developing more effective group recommender systems.

Figures (2)

• Figure 1: The deep neural network architecture for the CA-MCGRS using Multi-Head Attention Mechanism.
• Figure 2: Comparison of the performance of different models across training and validation sets. The top-left subplot shows the Training MSE for all models, while the top-right subplot depicts the Validation MSE. The bottom-left and bottom-right subplots illustrate the Training RMSE and Validation RMSE respectively, providing a view of each model's prediction error over the course of training epochs. Models with lower MSE and RMSE values demonstrate better accuracy in the recommendation task.