Data fusion, ensemble and clustering to improve the classification accuracy for the severity of road traffic accidents in Korea

Abstract

Increasing amount of road traffic in 1990s has drawn much attention in Korea due to its influence on safety problems. Various types of data analyses are done in order to analyze the relationship between the severity of road traffic accident and driving environmental factors based on traffic accident records. Accurate results of such accident data analysis can provide crucial information for road accident prevention policy. In this paper, we use various algorithms to improve the accuracy of individual classifiers for two categories of severity of road traffic accident. Individual classifiers used are neural network and decision tree. Mainly three different approaches are applied: classifier fusion based on the Dempster–Shafer algorithm, the Bayesian procedure and logistic model; data ensemble fusion based on arcing and bagging; and clustering based on the k-means algorithm. Our empirical study results indicate that a clustering based classification algorithm works best for road traffic accident classification in Korea.