Medical Insurance Suggestion Based On Age

K. Nitalaksheswara Rao; NALABOLU PRAVALLIKA; AVAGADDA SANYASI NAIDU NITIN; NEDURI VISHNU PAVAN KUMAR; AYNAMPUDI CHAITANYA VARMA; KOMMINENI JEEVAN SURYA

K. Nitalaksheswara Rao School of Technology, GITAM University
NALABOLU PRAVALLIKA School of Technology, GITAM University
AVAGADDA SANYASI NAIDU NITIN School of Technology, GITAM University,
NEDURI VISHNU PAVAN KUMAR School of Technology, GITAM University,
AYNAMPUDI CHAITANYA VARMA School of Technology, GITAM University,
KOMMINENI JEEVAN SURYA School of Technology, GITAM University,

Abstract

This project focuses on solving the complex issue of medical insurance cost estimation by leveragingpersonaldemographicandhealth-relatedfactors.Therisingcostofhealthcarehas ledto increased interest in accurate and personalized insurance pricing models, which can provide valuable insights to individuals when choosing the right insurance plan. This project uses adata-set containing key attributes such as age, sex, body mass index (BMI), smoking status, region of residence, number of children, and actual medical charges.

Topredict insurancechargeseffectively,weemployedvariousmachinelearningmodels, including LinearRegression,RidgeRegression,LassoRegression,DecisionTree,RandomForestRegression, and XGBoost. The purposeofusing multiple models was to compare their predictive performance andidentifythe most suitable modelforthistask.Each modelwasevaluatedusingkeymetrics like the R-squared score and Mean Squared Error (MSE) to determine its accuracy and reliability in predicting the cost of medical insurance.

Among the modelstested,the RandomForest Regression, after undergoing hyper parameter tuning, outperformed the other models in terms of both prediction accuracy and error minimization. This model’sabilitytohandlecomplex,non-linearrelationshipsbetweenthe featuresandtargetvariable (insurancecharges)madeittheidealchoiceforourproject.Hyper parametertuningfurtherenhanced its performance by optimizing key parameters, such as the number of trees in the forest and the maximum depth of each tree.

Byproviding individuals with accurate predictions oftheir insurance costs based on their personal data,this modelcan help people make more informed and cost-effective insurance decisions. This approach not only benefits users but also provides insurance companies with a toolto offer better- customized pricing, contributing to moreefficient and customer-friendlyhealth insurance systems.