Speed violation analysis of heavy vehicles on highways using spatial analysis and machine learning algorithms.

Area of Science:

• Transportation Engineering
• Traffic Safety
• Data Science

Background:

• Technological advancements enable higher vehicle speeds and improved road infrastructure, leading to increased instances of speeding.
• Speed violations by heavy vehicles pose significantly greater risks of life and property damage compared to light vehicles due to their carrying capacity.

Purpose of the Study:

• To analyze heavy vehicle speed violation data in Turkey.
• To classify provinces based on speed violation intensity using machine learning algorithms.
• To identify high-risk areas and propose targeted interventions for reducing speeding.

Main Methods:

• Data on heavy vehicle speed violations in Turkey were collected and categorized into 10 intensity classes.
• Support Vector Machines (SVM), Naive Bayes (NB), and K-Nearest Neighbors (KNN) algorithms were employed for provincial classification.
• Spatial analysis was utilized to create speed violation density maps based on the most accurate NB algorithm classification.

Main Results:

• The Naive Bayes (NB) algorithm demonstrated the highest accuracy in classifying speed violation rates among the tested algorithms.
• Density maps revealed that provinces such as Erzurum, Konya, and Muğla exhibit the highest rates of heavy vehicle speed violations.
• Further examination focused on heavy vehicle mobility in these identified high-violation provinces.

Conclusions:

• Implementing proposed measures can significantly mitigate material and moral damages caused by heavy vehicle speeding.
• Policymaker recommendations are crucial for reducing speed violations and enhancing road safety for heavy vehicles.
• Targeted interventions in high-violation areas like Erzurum, Konya, and Muğla are recommended to improve traffic safety.