Analyzing the Leading Causes of Traffic Fatalities Using XGBoost and Grid-Based Analysis: A City Management Perspective

Abstract

Traffic accidents have been one of the most important global public problems. It has caused a severe loss of human lives and property every year. Studying the influential factors of accidents can help find the reasons behind. This can facilitate the design of effective measures and policies to reduce the traffic fatality rate and improve road safety. However, most of the existing research either adopted methods based on linear assumption or neglected to further evaluate the spatial relationships. In this paper, we proposed a methodology framework based on XGBoost and grid analysis to spatially analyze the leading factors on traffic fatality in Los Angeles County. Characteristics of the collision, time and location, and environmental factors are considered. Results show that the proposed method has the best modeling performance compared with other commonly seen machine learning algorithms. Eight factors are found to have the leading impact on traffic fatality. Spatial relationships between the eight factors and the fatality rates within the Los Angeles County are further studied using the grid-based analysis in GIS. Specific suggestions on how to reduce the fatality rate and improve road safety are provided accordingly. INDEX TERMS XGBoost, factors analysis, GIS, grid-based analysis, non-linear machine learning, traffic fatality.

A Spatial Analysis Methodology Based on Lazy Ensembled Adaptive Associative Classifier and GIS For Examining the Influential Factors on Traffic Fatalities

“…These features have been partly discussed by previous literature [4], [10], [22], [46]. All of them exhibit explicit threats that lead to fatal accidents.…”

“…For example, VSAFETY2_W and VTYPE_3 are talking about two typical Vulnerable Road Users (VRUs), which are motorcycle drivers and pedestrians. Studies have shown that these VRUs have five times higher fatality rates than typical car-car accidents [46] because they have no protection equipment such as seat belts or airbags.…”

Time Series Analysis of Hemorrhagic Fever with Renal Syndrome in Mainland China by using XGBoost Forecasting Model

“…From studies in other fields, the XGBoost model performs well in predicting nonlinear time series [26][27][28][29]. By integrating multiple CART trees, the XGBoost model can acquire a better generalization performance than single one.…”

Prediction of Ambient PM2.5 Concentrations Using a Correlation Filtered Spatial-Temporal Long Short-Term Memory Model

“…Deep learning, also known as deep structured learning or hierarchical learning, is a class of machine learning that uses multiple layers of non-linear processing units for feature selection and data modeling [20]. Due to its powerful learning ability, deep learning has been applied to a number of fields such as computer vision, speech recognition, language processing [21,22], and has achieved state-of-the-art performance [23].Development of deep learning can be traced back to the machine learning technology. Due to the intrinsic linear assumption of traditional statistical methods, scholars tried to adopt non-linear machine learning methods to better fit the non-linear mechanism of real-world data like air pollution.…”

“…Deep learning, also known as deep structured learning or hierarchical learning, is a class of machine learning that uses multiple layers of non-linear processing units for feature selection and data modeling [20]. Due to its powerful learning ability, deep learning has been applied to a number of fields such as computer vision, speech recognition, language processing [21,22], and has achieved state-of-the-art performance [23].…”