Prediction of Occupational Risks Using an Adaptive Neural Fuzzy Inference System in AZARAB Company

Abstract

Background and Objective: Nowadays, none of the industries wants an accident to happen in their workplaces, and therefore, they use different tools to accomplish this aim. One of these tools is risk analysis, which is capable of identifying risks and inappropriate situations. Due to the importance of occupational risk prediction and injury reduction, this study was conducted to investigate occupational risk prediction using different neural network algorithms. Materials and Methods: This applied research was performed based on causal and survey approaches. Accordingly, a database of 119 incidents in 2018 was included in this study, which was sufficient reliable due to the high accuracy of the neural network algorithms in the database. The dynamic artificial neural network algorithm had the highest accuracy (76%) in predicting occupational injury. Results: Based on the results, the most important criteria affecting the risk of occupational injury were day-time, type of accident, and hazardous situations involved in the accident. Conclusion: This research can offer practical applications for Azarab company since this company can put all the vulnerabilities together, predict the risk of each of these situations by implementing the neural network algorithm, and accordingly take measures to provide risk control instructions.

Keywords

• accuracy
• most important features
• neural network
• occupational injury risk
• prediction