Healthcare Insurance Cost Prediction Using Machine Learning

Estimating healthcare insurance charges is a useful application of predictive analytics because recorded costs can vary with demographic and lifestyle characteristics. This paper presents a machine learning framework for estimating individual medical insurance charges from attributes such as age, sex, body mass index (BMI), number of dependent children, smoking status, and geographical region. The proposed workflow includes dataset inspection, data preprocessing, exploratory data analysis, feature engineering, categorical encoding, training of multiple regression algorithms, and evaluation on held-out data. Linear Regression, Ridge Regression, Decision Tree Regression, Random Forest Regression, and Gradient Boosting Regression are identified as candidate models for comparison. Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and the coefficient of determination (R²) are proposed as evaluation measures. The framework is intended to connect a fitted prediction pipeline with an application interface and analytical visualizations. The experimental evaluation has not yet been completed; therefore, this paper does not report model scores or identify a best-performing algorithm. The paper establishes a transparent and reproducible procedure for completing the comparison and reporting findings based on actual experiments.