Prognostic model when choosing pension saving strategies: selection algorithm and visualization of the results
Keywords:pension provision, CHAID algorithm, correspondence analysis, non-linear dependence, visualization of results of classification model
AbstractThe article is devoted to the elaboration of prognostic model of behavior; the study is based on the survey involving people of working age (men aged 18-60, women aged 18-55). It is shown that the CHAID algorithm is good for modeling non-linear dependence on the age for those who transferred the funded part of the pension from the state fund to the non-state fund on their age. Five age groups showing different retirement behavior were singled out. A special attention was paid to the visualization of the model obtained by correspondence analysis.
THEORY AND METHODOLOGY