Adapted Ridge Regression based on the Bayesian Likelihood Factor and its Application to a Model for Assessing the Impact of Health on the Well-being of Citizens of the Russian Federation
https://doi.org/10.25205/2542-0429-2024-24-2-99-134
Abstract
This article is aimed at their comparison of standard Ridge model based on default MSE evaluation of optimal lambda value with Baessyan criteria. Goal is to test adapted Ridge model on highly correlated data. Proposed method, unlike default one, has another advantage aside from conventional factors importance assessment – it give the way for p-value evaluation based on these Ridge coefficients. Those method gives a chance to tune parameters more deeply. In this case, there is a multidimensional sample in which it is necessary to filter out the least significant factors. At the same time, the data is characterized by homogeneity and various types of outliers. This justifies the choice of the specified methods. Data is based on the poll conducted among the population of the Russian Federation under the guidance of Higher School of Economics.
Supporting agencies: The article was prepared according to the research project «Regional and municipal strategic planning and management in the context of modernization of state regional policy and development of the digital economy»
About the Author
A. A. Zabolotsky
Institute of Economics and Industrial Engineering SB RAS
Russian Federation
Russian Federation
Alexey A. Zabolotsky - Candidate of Economic Sciences, Researcher, Department of Regional and Municipal Management, Institute of Economics and Industrial Engineering of the Siberian Branch of the Russian Academy of Sciences.
Novosibirsk
References
1. Cule E., Vineis P., De Iorio M. Significance testing in ridge regression for genetic data. BMC bioinformatics, 2011, vol. 12, № 1, pp. 1−15. DOI: 10.1186/1471-2105-12-372
2. Zhang S. et al. Kernel ridge regression for general noise model with its application. Neurocomputing, 2015, vol. 149, pp. 836−846.
3. Cawley G. C. et al. Heteroscedastic kernel ridge regression. Neurocomputing, 2004, vol. 57, pp. 105−124.
4. Burnaev E., Nazarov I. Conformalized kernel ridge regression. 15th IEEE international conference on machine learning and applications (ICMLA). IEEE, 2016, pp. 45−52.
5. Marquardt D. W., Snee R. D. Ridge regression in practice. The American Statistician, 1975, vol. 29, № 1, pp. 3−20.
6. Exterkate P. Model selection in kernel ridge regression. Computational Statistics & Data Analysis, 2013. vol. 68, pp. 1−16.
7. Garau M. et al. Is the link between health and wealth considered in decision making? Results from a qualitative study. International journal of technology assessment in health care, 2015, vol. 31, № 6, pp. 449−456.
8. Ghimir U. The impact of health on wealth: empirical evidence. Available at SSRN 3754401. 2020.
9. Shi X. The health-wealth nexus for the elderly: Evidence from the booming housing market in China. Labour Economics, 2022, vol. 78, p. 102247
10. Bristow P. Can Health Outcomes and Inequalities be Improved While Containing Costs? The Role of Voluntary Health Insurance in Universal Health Care Systems in the OECD. 2023.
11. Liu X. Q., Gao F. Linearized ridge regression estimator in linear regression. Communications in Statistics-Theory and Methods, 2011, vol. 40, № 12, pp. 2182−2192.
12. Kaneva M. A., Baidin V. B. Income Inequality and Self-Assessed Health in Russia. ЕСО, 2019, no. 12, pp. 105−123. DOI: 10.30680/ЕСО0131-7652-2019-12-105-123 (in Russ.)
Review
For citations:
Zabolotsky A.A. Adapted Ridge Regression based on the Bayesian Likelihood Factor and its Application to a Model for Assessing the Impact of Health on the Well-being of Citizens of the Russian Federation. World of Economics and Management. 2024;24(2):99-133. (In Russ.) https://doi.org/10.25205/2542-0429-2024-24-2-99-134
Views:
188
Email this article 
