|
KMID : 1124020210370020105
|
|
Korean Social Security Studies
2021 Volume.37 No. 2 p.105 ~ p.138
|
|
|
Developing a Prediction Model for Elderly Poverty and Deriving Important Rules Using Decision Tree: Focusing on the Characteristics of Current Elderly Generation
|
|
Kim Deok-Hyun
Jeong Dae-Yul
Yoo Dong-Hee
|
|
|
Abstract
|
|
|
|
The purpose of this study was to develop a prediction model for the elderly poverty using data from KLoSA (Korean Longitudinal Study of Ageing) and to derive important rules by analyzing the major factors that determine the elderly poverty. This study attempted to identify the differences by dividing the elderly poverty in terms of relative poverty and absolute poverty. To this end, we selected the elderly who are 65 years or older from the 7th respondents to KLoSA. Next, we developed a prediction model for the elderly poverty using a decision tree. The synthetic minority over-sampling technique (SMOTE) was used to alleviate the overfitting problem that may occur when developing the prediction model. The results of this study showed that absolute poverty shows a high correlation with private transfer income and relative poverty shows a high correlation with public transfer income. In addition, we derived seven rules for absolute poverty and eleven rules for relative poverty. Using the derived rules, we proposed meaningful strategies to alleviate absolute poverty and relative poverty.
|
|
|
KEYWORD
|
|
|
elderly poverty, relative poverty, absolute poverty, data mining, decision tree
|
|
|
FullTexts / Linksout information
|
|
|
|
|
|
Listed journal information
|
|
|