Ranking Highway Accident Black Spots Using Time Series Trend Analysis, Kernel Density Estimation, and Principal Component Analysis: A Case Study on Thai Highways

Abstract

This study proposes a method for identifying and ranking highway segments with high accident frequencies in Thailand, utilizing spatial and temporal analysis techniques. The Kernel Density Estimation (KDE) is used for spatial analysis, highlighting areas with high concentrations of accidents by estimating accident intensity across road segments. Additionally, Autoregressive Integrated Moving Average (ARIMA) models are applied to analyze accident trends from 2015 to 2022. The method helps identify patterns in accident frequency over time. These spatial and temporal factors are integrated through Principal Component Analysis (PCA), offering a comprehensive assessment of accident risks. This integration method ensures that the accident black spot identification considers both spatial and temporal factors.