This thesis examines insurance pricing with the goal of improving predictive accuracy through a comparative analysis of traditional actuarial techniques and modern machine learning algorithms. By utilizing real-world datasets from insurance companies, the research applies five distinct methodologies to analyze the key variables within the insurance dataset. The primary objective is to identify the most effective approaches to forecasting claim amounts. The findings of this study seek to advance predictive accuracy and provide substantial business value, thereby promoting innovation and excellence in risk management within the insurance industry.