Developing a Prediction Model to Identify Blood Types Most Susceptible to Viral Hepatitis Based on the CRISP-DM Methodology

General Background: Viral hepatitis is a prevalent disease worldwide, with hepatitis B and C posing significant public health challenges. While most cases resolve naturally, chronic infections contribute to severe complications. Specific Background: Genetic predisposition, including blood type, has been hypothesized as a risk factor for viral hepatitis; however, its role remains unclear. Knowledge Gap: Limited studies have analyzed the association between ABO blood groups and susceptibility to hepatitis B and C using machine learning techniques. Aims: This study aims to determine the blood groups most susceptible to hepatitis B and C by applying advanced machine learning models. Results: Using a dataset of 500 patients and CRISP-DM methodology, the findings indicate that blood type B has the highest susceptibility (38% infection rate), while type O shows the lowest risk (15%). Statistical analysis (Chi-square, p < 0.01) confirms a significant correlation between blood group B and hepatitis infection. The XG-Boost model achieved the highest predictive accuracy (91%), identifying blood type B as the second most influential risk factor after age. Novelty: This study provides empirical evidence linking genetic factors, particularly blood type B, with hepatitis susceptibility using data-driven approaches. Implications: The findings highlight the importance of blood type screening in high-risk populations and the necessity of targeted prevention strategies. Highlights: Blood type may influence susceptibility to hepatitis B and C. Blood type B shows highest risk; XG-Boost model achieves 91% accuracy. Blood type screening aids early detection and targeted prevention strategies. Keyword: Random Forest algorithm, Hepatitis B and C, KNN algorithm, Blood Groups, Decision Tree algorithm, support vector machine algorithm ,XG-Boost algorithm, neural network algorithm.