Research on online consumption preference behavior mining and recommendation based on BKAN model

Abstract

 This article aims to explore the in-depth mining of online consumption preference behavior and personalized recommendation strategies, and proposes a new method that integrates the BKAN (Bayesian Network Enhancement Based on Association Rules) model. The algorithm first uses association rule mining technology to identify consumption patterns, and then constructs a Bayesian network to enhance the model's ability to capture dynamic changes in consumption preferences. By introducing an attention mechanism to optimize node weight distribution, the BKAN model can more accurately predict users' future consumption intentions. Experimental results show that compared with traditional recommendation algorithms, the integrated BKAN model has significant advantages in improving recommendation accuracy and user satisfaction, providing strong support for personalized services on e-commerce platforms. This research provides a new perspective for online consumption preference analysis and helps promote the development of intelligent recommendation systems.