A Comparative Analysis of Different Machine Learning Techniques For Medical Insurance Premium Prediction

Abstract

Predicting health insurance premiums is a crucial task for both insurance companies and policyholders. This paper explores the use of regression approaches to predict health insurance premiums. The study uses a dataset of insurance premiums and related variables, including age, gender, BMI, and location to determine the factors that have the biggest effects on insurance rates. The study found that age and BMI were the main predictors of insurance rates, with gender and location having a relatively little influence. To create predictive models, the study combined various regression approaches, including logistic regression, Random Forest Regression, Decision Tree Regression, Gradient Boosting Regression, and linear regression. The predictive models were highly accurate, with a mean absolute error of less than 5% across all examined samples. The findings have significant implications for both insurers and customers, as accurate premium pricing is essential for insurers to maintain their financial health and offer customers coverage at reasonable prices. Accurate premium forecasts help policyholders plan their spending wisely for insurance costs and make informed decisions regarding their healthcare coverage.