<![CDATA[Employee stress has become a critical issue affecting organizational productivity, well-being, and performance, especially in dynamic work environments. This study proposes an integrated mobile-based stress prediction and recommendation system that combines Long Short-Term Memory (LSTM) and Neural Collaborative Filtering (NCF) to identify employee stress levels and provide personalized improvement recommendations. Experimental evaluation using 1000 datasets was used to test the LSTM and NCF models. The LSTM model was used to predict stress levels due to its ability to capture complex patterns in multidimensional data, while NCF was used to generate personalized recommendations based on collaborative patterns. The results showed that the LSTM model achieved superior classification performance with 98% accuracy and the recommendation evaluation showed good convergence performance, with a Hit Ratio reaching 0.92 and a Normalized Discounted Cumulative Gain (NDCG) reaching 0.89, indicating high recommendation relevance. Furthermore, the system usability evaluation using the System Usability Scale (SUS) involving 30 respondents resulted in an average score of 80.81, which is categorized as excellent usability. The integration of deep learning and collaborative filtering into a mobile platform provides an effective and intelligent solution for employee stress prediction and intervention. This study contributes to the development of an adaptive occupational health monitoring system and demonstrates the potential of AI-based mobile applications in supporting mental health management in the workplace.]]>
Copyrights © 2026
