Effective Risk Analysis and Predictive Modeling in Motor Insurance in Saudi Arabia

Accurate prediction of motor insurance premiums that correspond with actual claims are critical to the sustainability of insurance companies. However, predicting premiums is a challenging task due to the complexity of risk factors. This study aims to identify significant risk factors and develop predictive models for motor insurance pricing within the Saudi context, using real data obtained from one of the leading insurance providers in Saudi Arabia. The dataset consists of 71,280 records and 26 features of insurance claims reported during the year of 2023. After preprocessing the data, significant risk factors are identified using Analysis of Variance (ANOVA), which are used later to build the prediction models. The findings reveal that vehicle body type and manufacturing country emerged as the most influential risk factors. The evaluation metrics (R², MAE, MSE) have been applied to evaluate the best-performing machine-learning pricing prediction model (Decision tree, Neural network, Generalized linear model, and Random Forest). The results of our evaluation show that the Random Forest model consistently outperformed the other models in terms of prediction accuracy. The study contributes to motor insurance industry in Saudi Arabia by supporting informed risk assessment within the Saudi Takaful insurance operations. It highlights the performance of prediction models for motor insurance pricing in Saudi Arabia.