MedPredict Solutions: Medical Insurance Cost Prediction Using Machine Learning Algorithms

Authors- Assistant Professor Miss Shilpa Tripathi, Achintya Nivsarkar, Aditya Dubey, Ajay Shrivas, Vijay Shrivas

Abstract-This paper explores MedPredict Solutions, a machine learning system designed to predict health insurance prices more accurately and personally. By considering factors like age, BMI, smoking habits, pre-existing conditions etc, the model helps individuals get clearer estimates to better plan their finances. Insurance companies can also use this data-driven approach to price their products more fairly, leading to greater transparency in pricing. A number of machine learning models, namely Linear Regression, Decision Trees, Random Forest, and Gradient Boosting Machines, were adopted towards this prediction task. Each was tested on such performance metrics of Mean Squared Error and R- squared. Nonlinear models – Random Forest and GBM were the first indicators that performed far better than the more traditional linear models used because of their higher ability to identify the nonlinear relationships between health factors and insurance costs. The best result was achieved with the GBM model that was able to achieve the lowest MSE of 158.32 and R² score of 0.87, which signified the model’s ability to understand the intricacies characteristic of health care data. Beyond the insurance pricing problem, the present study may have wide applicability. For the individual, it offers the ability to make better financial planning based on personal health risks. It promotes fairness and transparency in pricing for insurers, while providing data-driven insights that can improve healthcare policies and lead to more equitable financing of the healthcare system.

DOI: /10.61463/ijset.vol.12.issue5.249