A Machine Learning Approach to Comparing Different Regression Models to Predict Bangladesh Life Expectancy using Multiple Depend Features

Abstract

The average longevity of a person is measured by life expectancy. The length of one's life is determined by a number of factors. We utilized GDP, rural population growth, urban population growth, services value, industry value, food production, permanent cropland, cereal production, agriculture, forestry, and fisheries value as indicators of life expectancy. We may examine all of the factors that influence life expectancy, such as the negative link between life expectancy and rural population. We can observe how personality traits are linked to life expectancy and impact how we spend our lives. To evaluate which regression models are the most accurate, we use eight different regression models. The Extreme Gradient Boosting Regressor has the greatest accuracy and the least error of all the models. It was 99 percent correct. K-Neighbors, Random Forest, and Stacking Regressor were all 94 percent accurate. Slightly Stacking was the most accurate of the bunch. We used K-Neighbors, Gradient Boosting, and Random Forest Regressor for the Stacking Regressor, and Random Forest for the meta regressor. Decision Tree has the lowest accuracy of all the models, at 79 percent. The Gradient Boosting Regressor comes in second with 96 percent accuracy. Multiple Linear Regression and Light Gradient Boosting Machine Regressor scored 88 percent and 87 percent, respectively. This study assists a country in enhancing the value of its characteristics in terms of life expectancy.