Classification and prediction of accident factors in the firework industries

Many fire accidents have occurred in recent times, and almost 95% of fire industry accidents are caused by manual errors and frictions. Other factors that have a large impact on casualties include temperature mismatching, human carelessness, chemical frictions, static electricity, and atmospheric conditions. These causes create massive destruction in the fireworks industry and have led to a high death rate in the past decades. This paper introduces a new machine learning strategy called Extensive Fire Based Risk Factor Estimation (EFBRFE) based on a past accident dataset. It analyzes the factors influencing an explosion in the firework industry related to a fatal accident. The proposed approach’s results demonstrate the efficacy of a new model, the level of risk assessment, and the identification of human flaws in the working environment required to lead a safe manufacturing scenario. These significant issues provide a practical and intellectual framework for the Virudhunagar district firework manufacturing industry. The dataset pertains to fireworks accidents in this district. The accuracy, sensitivity, and specificity of this model are 99.10%, 98.4%, and 97.6%, respectively. This model is used intelligently to classify accidental factors and predict risk and consequence levels.