Human Life Expectancy Prediction Through Machine Learning Models: A Comparative Analysis of Classification and Regression Models
https://doi.org/10.5281/zenodo.19677030
Keywords:
machine learning, XGBoost, Random Forest, Life expectancy prediction, healthcare analytics, regression models, classification models, public healthAbstract
Life expectancy is a very important issue of economic development and public health. The correct measure of life expectancy problems can significantly help in planning and policy-making decisions related to the population’s healthcare. In this proposed research, various machine learning models of Regression and classification have been implemented to predict the life expectancy and country development status. This research has used a dataset of 2,938 samples having 22 features, including economic, demographic, and health indicators. The preprocessing techniques were applied for data analysis and cleaning purposes. This study implements the 8 machine learning regression models (Gradient Boosting, Random Forest, Ridge, Linear Regression, Lasso, XGBoost, Decision Tree, and SVR), and 10 Classification models, including SVM, KNN, CatBoost, Naïve Bayes, XGBoost, Gradient Boosting, LightGBM, Random Forest, Decision Tree, and Logistic Regression, were trained on the health dataset. The experimental results of this research reveal that the regression model XGBoost achieved the highest R2 score of 0.89 with an RMSE value of 3.25 years, while the classifier XGBoost achieved 96.8 % accuracy in predicting the country development status. This research highlights the effectiveness of multiple machine learning models for task prediction in the public health domain.
