|
Forecasting health insurance premium using machine learning approaches |
|---|---|
| รหัสดีโอไอ | |
| Creator | Samir K. Bandyopadhyay |
| Title | Forecasting health insurance premium using machine learning approaches |
| Contributor | Shawni Dutta, Payal Bose |
| Publisher | Asia-Pacific Journal of Science and Technology |
| Publication Year | 2566 |
| Journal Title | Asia-Pacific Journal of Science and Technology |
| Journal Vol. | 28 |
| Journal No. | 6 |
| Page no. | 13 |
| Keyword | Ensemble approach, Feature importance, Health insurance price, Machine learning, Random forest, Regression analysis |
| URL Website | https://www.tci-thaijo.org/index.php/APST |
| Website title | https://so01.tci-thaijo.org/index.php/APST/article/view/253968 |
| ISSN | 2539-6293 |
| Abstract | A medical emergency can impact anybody at any moment and have a significant psychological and economic consequence. Health insurance encompasses different costs including hospital charges, medicine costs, physician consultation fees, etc. The importance of health insurance cannot be underestimated, given the exponential rise in healthcare expenses. This research has attempted to forecast the cost of health insurance premiums. To address this problem, an automated system can be built that analyzes an individual's health complications and forecasts associated costs. Machine learning-based techniques can be employed to design the automated system. This research work will use ensemble-based machine learning approaches to evaluate an individual's risk and anticipate premium costs. This will allow the insurance firms to set a minimal fee with a higher profit in order to attract more policyholders. Random Forest, CatBoost, Extra Trees, AdaBoost, Extreme Gradient Boost, and Gradient Boost models are popular ensemble-based algorithms that are applied to an individual's health information and used to estimate premium price. The automated model can be developed by applying the Random Forest model with a Mean Square Error (MSE) of 0.01 and Mean Absolute Error (MAE) of 0.0394, according to the comparative study of these employed models. The graphical representation of the comparative analysis is depicted in Figure 1A and 1B based on MAE and MSE respectively. The research finds relative ranking among the interfering factors for determining medical cost for an individual. According to the Random Forest model's findings, a person's age has the greatest impact on determining the related premium. |
