Health Insurance Claim Prediction and Risk Assessment Using an AI-Based Machine Learning Framework

Abstract

The rising prevalence of health issues and increasing medical expenditures have
made health insurance an essential component of modern society. Insurance providers such as
LIC, ICICI, HDFC ERGO, and Star Health offer financial support to individuals by covering
medical expenses; however, accurately estimating claim amounts remains a challenging task.
This study utilizes a comprehensive dataset comprising generalized, hospitalization, and claimrelated
data, where each record reflects an individual’s insurance charges along with
demographic and health attributes. Machine learning (ML) techniques are employed to analyse
patterns in hospitalization costs and insurance payments, enabling effective grouping and
prediction of claim amounts. In the contemporary insurance domain, data-driven approaches play
a crucial role in risk assessment and financial planning. ML models are capable of extracting
complex, non-linear relationships from historical data, including patient demographics, treatment
details, and past claims. This study proposes a computational intelligence-based framework
using algorithms such as Random Forest and Support Vector Machines (SVM) to predict health
insurance expenses. Experimental results demonstrate that Random Forest achieves the highest
prediction accuracy of 90%, outperforming SVM with 83%. The comparative analysis highlights
the effectiveness of ensemble learning methods in improving predictive performance. The
proposed approach supports better decision-making, risk management, and resource allocation in
the healthcare insurance sector.