Kamu Yönetimi Açısından Yapay Zekanın Türkiye’deki Bütçe Açıklarının Tahmini Üzerine Etkisi
The Effect of Artificial Intelligence on Estimation of Turkey’s Budget Deficits from the View of Public Administration
Author(s): Aylin Konu, Ahmet Yılmaz AtaSubject(s): Economic history, Economic policy, Post-War period (1950 - 1989), Transformation Period (1990 - 2010), Present Times (2010 - today), Fiscal Politics / Budgeting, ICT Information and Communications Technologies
Published by: Ahmet Arif Eren
Keywords: Public Administration; Budget Deficit; Artificial Intelligence; Fuzzy Neural Networks;
Summary/Abstract: Technology progress and digitalization have led to serious transformations in many areas. These transformations have also affected the understanding of public administration and added a new dimension to research in this area. In this context, the interaction of budgeting and budget deficits with digitalization and artificial intelligence will constitute the study area of this study. Ensuring budget balance is one of the most important macroeconomic targets for countries. It is stated that, factors such as high inflation, external payment difficulties, insufficient savings rates and increases in public debt are effective in budget deficits. In this framework, ensuring fiscal discipline is considered a necesssery condition for achieving sustainable economic growth. In this study, Turkey’s budget balance, different from the other classical methods used in the literature, will be tested using a fuzzy neural network, one of the sub-branches of artificial intelligence. In this framework, the budget balance will be estimated using the factors affecting the budget deficits in the period of 1980-2019. The validity of the Fuzzy Neural Networks method in this estimation will also be investigated. As a result, the proposed NFS model can be used successfully to estimate the budget deficit since obtained error rate between the estimation results and the actual budget deficit is minimal.
Journal: Fiscaoeconomia
- Issue Year: 6/2022
- Issue No: 2
- Page Range: 636-655
- Page Count: 20
- Language: Turkish