Medical health insurance price prediction using supervised machine learning

Author

Makore, Lovemore.

Title

Medical health insurance price prediction using supervised machine learning

Abstract

This research explores the use of supervised machine learning, particularly random forest algorithms, to forecast medical health insurance prices in Zimbabwe, addressing the challenges posed by high healthcare costs and limited access to affordable insurance. A dataset encompassing demographic and health-related records from medical insurance was collected and analyzed. The random forest model developed on this dataset was evaluated using key performance metrics: Mean Absolute Error (MAE) of 0.477, Mean Squared Error (MSE) of 1.249, Root Mean Squared Error (RMSE) of 1.119, R-squared (R²) value of 0.883, and Mean Absolute Percentage Error (MAPE) of 1.837%. These metrics collectively assess the accuracy and reliability of the model in predicting insurance prices. The findings indicate that the random forest approach achieves high predictive accuracy, showcasing its potential to enhance pricing transparency, improve risk assessment practices, and elevate customer satisfaction within Zimbabwe's healthcare insurance landscape. Moreover, the study identifies critical determinants influencing insurance costs, such as age, smoking status, and BMI, underscoring opportunities for innovative applications in personalized insurance pricing strategies, health risk evaluations, and advancements in Insurtech solutions.

Date

June 2024

Publisher

BUSE

Keywords

Machine learning,

Medical health insurance,

Health risk evaluations,

Supervisor

Mr. W. Kanyongo,