Road traffic accident case status prediction integrating a modified c4.5 algorithm

Abstract

C4.5Algorithm is one of the most popular algorithms for rule base classification. The problem C4.5 algorithms is dealing with continuous attribute and the accuracy of prediction is affected and the inducer will choose an over fitted classifier. In this study, there are 25 variables and 799 accidents record generated from this tree. The case status is the best one that can be used for prediction. based on the total training set of instances are increases the accuracy of 69.83% while the incorrect classified instances are 30.16%. the precision solved 725 while the unsolved is 663, recall 746 solved while the unsolved is 638 and F-measure of the solved case status is 735 while the unsolved is 650. The accuracy of Case status of the Road Traffic Accident are increases. The confusion matrix classified as the solve is 224 while the unsolved is 114. Based on the result the common causes of the accident are that Suspect use alcohol with the 615 records which the highest number record with heavy vehicle in the afternoon and evening during rainy and dry season. The occurrence of accident is in Agusan del Norte with 328 accidents.