Work Accident Analysis with Machine Learning Techniques

Abstract

All over the world, serious investments have been made in recent years on workers' health and safety. With the importance given to health and safety of workers, new studies have been performed. In this study, data mining and machine learning techniques are applied to the real worker accident data. Firstly, data cleaning and feature selection are performed to use machine learning algorithms, then the classification result obtained by using K-nearest neighbors (KNN) and Naive Bayes (NB) classification algorithms. Accuracy and F-measure metrics were used to measure classification success. The highest success rate was obtained with the KNN algorithm by 10 cross-validation. These values are 0.994075 and 0.993257 for the accuracy and F measure respectively.