Vehicle Accident Severity Rules Mining Using Fuzzy Granular Decision Tree

Abstract

Road accident is a disaster that vocalizes a major cause of disability, untimely death and the loss of human lives. Therefore, investigating the condition of road accidents for prediction and prevention purposes on highways is significant. In this paper, we propose a new fuzzy granular decision tree to generate the road accident rules applying the discrete and continuous data stored in accident databases. Among all critical factors in the occurrence of traffic accidents, environmental factors and road design (geometry) are considered in this study. This method establishes an optimized fuzzy granular decision tree with the minimum redundancy and road accident severity classification using fuzzy reasoning. California highways were considered as the case study to examine the proposed approach. The experimental results demonstrate that the proposed method is approximately 16% more accurate than the fuzzy ID3 method with less redundancy in constructing the decision tree.