Back to Search Start Over

Research on Forecast Model of Elderly Pension Expenses Based on Data Mining Algorithm

Research on Forecast Model of Elderly Pension Expenses Based on Data Mining Algorithm

Authors :
Bin Zhu
Li Zhang
Yonggang Zhang
Source :
Journal of Physics: Conference Series. 1852:032051
Publication Year :
2021
Publisher :
IOP Publishing, 2021.

Abstract

With the intensification of the aging of our country’s population, research on how to better protect the living standards of the elderly, reduce the burden of the working population to support the elderly, and ensure the orderly and healthy development of the society, is currently the focus of attention. In order to prevent aging from bringing further negative impacts on our society or economy, this paper uses data mining algorithms as the research basis and uses time series models to make effective predictions. Therefore, this article first decomposes these data by analyzing the characteristics of elderly consumption data. The decomposed sequence has short correlation characteristics, and the prediction accuracy is significantly higher than that of the traditional time series model. However, this kind of time series model can only predict some offline data. In order to make it can be used for dynamic data prediction, this paper uses RBFNN online prediction algorithm. The RBF neural network is partially improved, and the parameters are calculated by improving the later SGD algorithm. The experimental conclusion shows that the prediction accuracy and efficiency of this algorithm are obviously more accurate than the neighbor clustering online training algorithm, and it achieves more effective online prediction. It can be concluded that the role of the elderly social security budget in economic management can fully reflect the income and expenditure of the elderly social security funds, improve the efficiency of fund use, and establish a unified national social security budget, thereby reducing the impact of aging on society Negative impact.

Details

ISSN :
17426596 and 17426588
Volume :
1852
Database :
OpenAIRE
Journal :
Journal of Physics: Conference Series
Accession number :
edsair.doi...........773a230c37e9f77dee58a63bd7fcab58
Full Text :
https://doi.org/10.1088/1742-6596/1852/3/032051