Implementation of Medical Insurance Price Prediction System using Regression Algorithms

• Main Authors: Vijayalakshmi, V., Selvakumar, A., Panimalar, K.
• Format: Conference Proceeding
• Language: English
• Subjects: Costs; Evaluation Mettices; Insurance; Manuals; Measurement uncertainty; Medical Insurance Cost; Neural networks; Prediction; Programming; Regression Algonthms; Support vector machines
• ISSN: 2832-3017
• DOI: 10.1109/ICSSIT55814.2023.10060926

Medical expenses are increasing day by day due to unexpected epidemic diseases. People have awareness about the medical insurance and claiming processes which is helping in critical situations. Predicting the medical insurance cost is very important work in health care sectors. Many researcheis applied different machine learning algoiithms to predict the insurance premium in Kaggle data set with seven attributes such as age, sex, bmi, children, smoker, region and charges. But these features only not suitable for prediction. In this paper, the dataset with 24 features including all relevant attributes needed for prediction of insurance cost was used. The implementation done using regression algorithms such as Linear Regression, Decision Tree Regression, Lasso Regression, Ridge Regression, Random Forest Regression, ElasticNet Regression, Support Vector Regression, K Nearest Neighbor Regression and Neural Network Regression in R Programming. There were seven metiices applied to measure the performance of the model namely Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), RSquared (R 2 ), Adjusted R-Squared (Adj. \mathrm{R}^{2}), and Explained Variance Score (EVS). Random Forest Regression outperformed with 0.9533 as RS quared value. This prediction system will reduce the manual work and give the prediction accurately in medical field.