Year: 2025
Type: article
Abstract: The rising complexity of healthcare expenses has increased the demand for accurate medical insurance cost prediction systems. This project presents a machine learning-based approach to estimate individual insurance charges by analyzing demographic and health-related features such as age, gender, body mass index (BMI), number of dependents, smoking habits, and residential region. A publicly available dataset was preprocessed by encoding categorical attributes, scaling numerical variables, and visualizing relationships between predictors and charges. Multiple regression and ensemble algorithms, including Linear Regression, Support Vector Regression, Ridge Regression, and Random Forest Regressor, were implemented and evaluated. Comparative analysis revealed that the Random Forest Regressor outperformed other models, achieving an accuracy of approximately 86% in predicting insurance premiums. To enhance usability, the trained model was deployed using a Flask web application that provides real-time predictions based on user inputs. The system not only estimates the expected cost but also categorizes the user’s BMI and assigns a risk level according to the predicted premium, thereby improving interpretability for end-users. Furthermore, predictions are logged systematically for monitoring and analysis, ensuring transparency of model behaviour. The integration of machine learning algorithms with a lightweight web framework demonstrates the feasibility of delivering practical and user-friendly solutions in the domain of health insurance analytics. This research highlights the significance of predictive modeling in fostering fair insurance pricing, assisting individual in financial planning, and aiding insurance providers in risk assessment.
Authors: Harshitha G J, Manjunath Kumar B H -
Language: English
Subfield: Health Information Management
Field: Health Professions
Domain: Health Sciences
Open Access status: hybrid
OpenAlex ID: W4414379205
FWCI:
Cites:
Cited by:
Related to: