Potent Real-Time Recommendations Using Multimodel Contextual Reinforcement Learning

Abstract

Widespread digitalization has led to almost all utilities and services thrive on an online medium. A real-time, personalized, and trend grasping recommendation system is necessary to enhance user experience and boost business on E-commerce platforms. We propose the Multimodel Contextual Reinforcement Learning (MMCR) constituting three novel features for real-time and customized recommendations. The first feature is user-item interactive state embedding which uses not only item information but also assigns weightage to this information according to its usage history. It gives higher importance to the newly clicked items by the users than the older ones. Second, we devised Contextual Cluster Exploration (CCE) strategy. This strategy enhances the item-choice recommendations by consistently reducing the randomness during exploration. The third novelty is an item-based multi-agent framework that can tackle the case of sparsely chosen items. Generally, such items are disregarded in a single agent model as the more popular items take supremacy. Our technique ensures that the user-item history per item is learned separately; thus, no item is neglected. MMCR has shown an average of 5% increase in CTR rate. Moreover, CCE exploration gives a considerably higher score than state-of-the-art exploration strategies. Thorough experimentation demonstrates that our proposed strategy has shown significantly improved results over various state-of-the-art strategies.