EATR: Emotion-Aware Temporal Reinforcement Learning Datasets for Multi-Device Recommendation

This repository contains datasets used in the research article“Emotion-Aware Temporal Reinforcement Learning for Serendipitous Recommendations in Multi-Device Environments”. The datasets include:(i) a synthetic emotion-aware multi-device interaction dataset designed for affective and temporal reinforcement learning,(ii) anonymized and aggregated cross-device temporal interaction logs representing realistic smart-environment usage patterns. All datasets are fully anonymized, contain no personally identifiable information, and are released to support reproducibility and future research in affect-aware recommender systems.