CLASSIFICATION OF FACTORS THAT INFLUENCE THE SEVERITY OF WORKER’S INJURIES USING THE CART METHOD (CLASSIFICATION AND REGRESSION TREE)

Abstract

ABSTRACT

PT Varia Usaha Beton is a company engaged in the manufacture of ready-made concrete, pre-cast concrete, concrete mansory, and broken stone. This company is a subsidiary of PT Semen Indonesia. In the production process, many cases of work accidents occur to workers. The high level of accidents that occur can disrupt the production process and cause the company's production to be less than optimal. Based on this urgency, this study was conducted to make predictions on the severity of injuries in the event of workplace accidents on workers using work accident data from 2016-2019, as many as 129 cases. The data used was obtained from the K3 unit at PT Varia Usaha Beton. Classification are made using a data mining approach with the decision tree method. The CART (Classification and Regression Tree) method is a type of decision tree used to make predictions. The results of this study are in the form of influential factors related to the severity of injury in the event of workplace accidents on workers which results in an accuracy rate of 81.1%. There are 3 factors that affect the severity of workers' injuries in the event of a work accident, where of the 3 factors there is 1 factor that is most influential on the occurrence of work accidents, namely work shifts. Two other influential factors are the age and location of the accident. CART method can be used in predicting the severity of injuries if work accidents occur in the concrete manufacturing industry. Thus the severity of injuries due to workplace accidents can be reduced by using a classification model for prevention and training in the company.

Key Words : Classification of Influential Factors, Severity of Injury in Workers, CART (Classification and Regression Tree) Method, Decision Tree.