Modeling of the Safety Climate and the Cultural Attitudes to Predict Unsafe Behaviors Using the Neuro-Fuzzy Inference System (ANFIS)

Background: Unsafe behavior in industries can be due to different factors. The aim of this study was to predict and model unsafe behavior using a safety atmosphere and cultural attitudes questionnaires. Methods: This study was a descriptiveanalytic and crosssectional examination that analyzed the data and predicted the unsafe behaviors of 90 construction workers using NeuroFuzzy Inference System (ANFIS) in MATLAB R2016a software. Results: In this study, the model of the safety atmosphere unsafe behavior and the model of the cultural attitudes unsafe behavior had the regression coefficients of 0.93373 and 0.9234, respectively. It showed that each of the parameters has a close relationship to the rate of the unsafe behavior. In this regard, a combination of the safety atmosphere and safety attitude parameters for the estimation of the unsafe behaviors achieved the better results with a regression coefficient of 0.9453 which indicates the direct effect of both parameters simultaneously on unsafe behavior. Conclusion: Based on the findings, it can be concluded that the neurofuzzy model can be used as an appropriate tool for predicting unsafe behavior in the industries.