Stress Relieving and Emotion Alleviating Activities Recommendation System Based on Deep Reinforcement Learning - A Novel Method Providing Slate Recommendations Which Utilizes SlateQ with Multi-Objective Optimization

This paper presents a novel Stress-Relief and Emotion-Alleviation Activity Recommendation System. The recommendation problem is framed as a sequential decision problem; thus, a Deep Reinforcement Learning (DRL) approach is used to implement it. The study adopts SlateQ – an algorithm developed by Google researchers for generating recommendation slates. Through the integration of Multi-Objective Reinforcement Learning (MORL) with a weighted sum strategy, the research aims to concurrently enhance stress reduction and emotion alleviation. The study extends the existing landscape of recommendation systems by combining DRL, MORL, and SlateQ to address the intricacies of stress and emotion management through personalized activity recommendations. This recommendation system is a part of the desktop application we developed, named "DevRelax," which helps improve the productivity of employees in IT companies. This research contributes to the advancement of intelligent systems designed to ameliorate emotional well-being and provides a comprehensive framework applicable in various contexts.