Investigation of Contributing Factors to Road Traffic Accidents in Thailand by Using Latent Class Analysis

Abstract

Traffic accident is a leading cause of death in Thailand. Understanding the patterns of accident helps determine the measures. Hence, Latent class clustering (LCC) analysis was conducted on 41,489 traffic road accident cases from the Transport Accident Management Systems (TRAMS) provided by the Ministry of Transport between 2021 and 2022 to show the patterns of traffic accidents. The analysis, which included six variables: road type, collision type, vehicle group, weather, time, and presumed cause, revealed four and five clusters of road accident patterns in the 2021 and 2022 datasets, respectively. Four comparable pairs of clusters are matched in those years. In addition, the vehicle group analysis was performed by LCC, giving similar results for six vehicle groups in both years. Vehicle group analysis showed the effect of vehicle variables on the characteristics of accident patterns as cluster to population ratio (C/P). The results suggest road safety policies should focus on the cluster of run-off-roadway accidents on straight sections caused by speeding. The LCC analysis also provides an advantage in the further application and policy evaluation, as progress can be tracked by the change in clustering behavior of the future accidents dataset.