Motorways are typically the safest road environment in terms of injury crashes per million vehicle kilometres; however, given the high severity of crashes occurring therein, there is still space for road safety improvements. The objective of this study is to compare the classification performance of five machine learning techniques for predictions of crash risk levels of motorway segments. To that end, data on crash risk levels, driving behaviour metrics, and road geometry characteristics of 668 motorway segments were exploited. The utilized dataset was divided into training and test subsets, with a proportion of 75% and 25%, respectively. The training subset was used to train the models, whereas the test subset was used for the evaluation of their performance. The response variable of the models was the crash risk level of the considered motorway segments, while the predictors were various road design characteristics and naturalistic driving behaviour metrics. The techniques considered were Logistic Regression, Decision Tree, Random Forest, Support Vector Machine, and K-Nearest Neighbours. Among the five techniques, the Random Forest model achieved the best classification performance (overall accuracy: 89.3%, macro-averaged precision: 89.0%, macro-averaged recall: 88.4%, macro-averaged F1 score: 88.6%). Moreover, the Shapley additive explanations were calculated in order to assist with the interpretation of the model’s outcomes. The findings of this study are particularly useful as the Random Forest model could be used as a highly promising proactive road safety tool for identifying potentially hazardous motorway segments.

High quality data on road crashes, road design characteristics, and traffic are typically required to predict crash frequency. Surrogate Safety Measures (SSMs) are an alternative category of indicators that can be used in road safety analyses in order to quantify various unsafe traffic events. The objective of this research is to exploit road geometry data and SSMs toward various road crash investigations in motorway segments. To that end, for this analysis, a database containing data on injury and property-damage-only crashes, road design characteristics, and SSMs of 668 segments was compiled and utilized. The results of the developed negative binomial regression model revealed that crash frequency is positively correlated with the average annual daily traffic volume, the length of the segment, harsh accelerations, and harsh braking. Moreover, four distinct clusters representing crash risk levels of the examined segments emerged from the hierarchical clustering procedure, ranging from more risk-prone, potentially unsafe locations to more safe locations. These four clusters also formed the response variable classes of a random forest model. This classification model used various road geometry data and SSMs as predictors and achieved high classification performance for all classes, averaging more than 88% correct classification rates.