Hotelling T2 Control Chart for Detecting Changes in Mortality Models Based on Machine-Learning Decision Tree

Mortality modelling is a practical method for the government and various fields to obtain picture of mortality up specific age particular year. However, some information on phenomenon may remain in residual vector be unrevealed from models. We handle this issue by employing multivariate control chart discover substantial cohort changes behavior that models still need address. The Hotelling T2 applied externally studentized deviance model, which already optimized using machine-learning decision tree. This study shows model with lowest MSE, MAPE, deviance, accomplishing simulations countries. In addition, more sensitive detecting signals singled out so we can perform decomposition determine attributes death outlying group case uses data country Saudi Arabia. overall results demonstrate our processing producing machine learning solution developing countries or limited produce accurate predictions through monitoring charts.