Predictive Analytics for Medical Insurance Premiums

This paper enables a machine leaning method to project the health insurance costs, ever more crucial in view of growing medical expense complexity. Both payers and buyers are served by stable insurance price estimates. The authors used a sample set consisting of medical history, demographics, and other contributing parameters to construct model predictions by invoking multiple machine learning methods, from ensemble modeling and regression.The models were assessed with measures such as R-Squared and Mean Absolute Error, showing the methods were able to accurately predict insurance premiums. The innovation provides useful insights for insurance practitioners, with the potential to maximize pricing strategies, improve risk estimates, and enhances decision-making in the health insurance market. The research also highlights potential of the machine learning to forecast medical insurance premiums using personal attributes-gender, age, BMI number, dependents, smoking habits ,and geographical location are collected and prepared. The datasets have been separated into two groups: training and testing the Artificial Neural Network model is developed and trained on TensorFlow and Keras, and it produced a high R-Squared value indicating its precision. The study found significant factors that determine insurance premiums. Generally, the results indicate that machine learning has great potential in predicting healthcare expenditures. Through these predictive models, insurance firms may provide more customized insurance products, simplify the underwriting process, and enable customers to make well known decisions about their health plans. Finally, this study seeks to strengthen Utilization of information for making decisions in the insurance market, to the advantage of both providers and policyholders.