CReSHARP: Cement Recommendation System for Health Risk Analysis and Prevention for Workers in Cement Industries

Abstract

Cement production is an essential industry, but it poses significant health risks to workers due to exposure to harmful dust and chemicals. This research introduces the Cement Recommendation System for Health Risk Analysis and Prevention (CReSHARP), a tool designed to mitigate these occupational hazards. CReSHARP utilizes a deep learning-based content recommender system to suggest preventive measures tailored to individual workers' symptoms and diagnosed diseases. The system's development involved creating a comprehensive dataset from literature reviews and employing a Generative Adversarial Network (GAN) to enhance data volume. The model was trained using k-Nearest Neighbors (k-NN), Support Vector Machine (SVM), and a hybrid CNN-LSTM architecture. Results showed the CNN-LSTM model achieved the highest accuracy in recommending preventive measures. The system's implementation can significantly improve worker safety by providing personalized health recommendations, thereby reducing the incidence of occupational diseases in the cement industry. Future work should focus on incorporating real-world data and continuously updating the system to enhance its applicability and reliability.