Comparative Forecasting of Some Key Economic Indicators Using Artificial Neural Networks and Ordinary Differential Equations: A Case Study of the Turkish Economy

Bahatdin Daşbaşı; Murat Taşyürek

Makale detayları

Comparative Forecasting of Some Key Economic Indicators Using Artificial Neural Networks and Ordinary Differential Equations: A Case Study of the Turkish Economy

Yazarlar:

Bahatdin Daşbaşı | Murat Taşyürek

Yayıncı:

Çanakkale Onsekiz Mart University

Ana Başlıklar:

Information Systems Development Methodologies and Practice

Dergi:

Journal of Advanced Research in Natural and Applied Sciences Vol.10 Issue 4, pp.922-937

Belge Tipi:

Makale

Yayın Tarihi:

12/2024

Belge Dili:

İngilizce

Document ID (Org No.):

2001338207

pISSN:

2757-5195

Anahtar kelimeler:

Özet:

 This study explores the relationships between the USD opening exchange rate, the annual change rate of the Consumer Price Index CPI, the housing loan interest rate in Turkish lira, and the residential construction cost index from January 2015 to May 2024 using data from the Turkish Statistical Institute TUIK. Artificial Neural Networks ANN and Ordinary Differential Equations ODE methods were employed to model the interactions among these four variables. In the ANN approach, each variable was modeled as the dependent variable in turn, with the remaining three serving as independent variables, resulting in four distinct analyses. The ODE model, on the other hand, provided a holistic analysis by capturing the time-dependent relationships among all four variables simultaneously. The ANN model predictions achieved accuracy rates of 87.2 for the USD opening exchange rate, 91.4 for the CPI annual change rate, 85.9 for the housing loan interest rate, and 93.1 for the construction cost index. Meanwhile, the ODE model demonstrated its strength by offering a more comprehensive framework with an overall accuracy of 94.6, effectively capturing the complex interdependencies among the variables. These findings highlight the strengths of both approaches: while the ANN model excels in analyzing individual variables, the ODE model offers a broader perspective by integrating all variables into a unified framework. This study contributes to developing economic forecasting models and provides valuable insights for decision-makers, particularly in times of economic uncertainty.

 This study explores the relationships between the USD opening exchange rate, the annual change rate of the Consumer Price Index CPI, the housing loan interest rate in Turkish lira, and the residential construction cost index from January 2015 to May 2024 using data from the Turkish Statistical Institute TUIK. Artificial Neural Networks ANN and Ordinary Differential Equations ODE methods were employed to model the interactions among these four variables. In the ANN approach, each variable was modeled as the dependent variable in turn, with the remaining three serving as independent variables, resulting in four distinct analyses. The ODE model, on the other hand, provided a holistic analysis by capturing the time-dependent relationships among all four variables simultaneously. The ANN model predictions achieved accuracy rates of 87.2 for the USD opening exchange rate, 91.4 for the CPI annual change rate, 85.9 for the housing loan interest rate, and 93.1 for the construction cost index. Meanwhile, the ODE model demonstrated its strength by offering a more comprehensive framework with an overall accuracy of 94.6, effectively capturing the complex interdependencies among the variables. These findings highlight the strengths of both approaches: while the ANN model excels in analyzing individual variables, the ODE model offers a broader perspective by integrating all variables into a unified framework. This study contributes to developing economic forecasting models and provides valuable insights for decision-makers, particularly in times of economic uncertainty.

 This study explores the relationships between the USD opening exchange rate, the annual change rate of the Consumer Price Index CPI, the housing loan interest rate in Turkish lira, and the residential construction cost index from January 2015 to May 2024 using data from the Turkish Statistical Institute TUIK. Artificial Neural Networks ANN and Ordinary Differential Equations ODE methods were employed to model the interactions among these four variables. In the ANN approach, each variable was modeled as the dependent variable in turn, with the remaining three serving as independent variables, resulting in four distinct analyses. The ODE model, on the other hand, provided a holistic analysis by capturing the time-dependent relationships among all four variables simultaneously. The ANN model predictions achieved accuracy rates of 87.2 for the USD opening exchange rate, 91.4 for the CPI annual change rate, 85.9 for the housing loan interest rate, and 93.1 for the construction cost index. Meanwhile, the ODE model demonstrated its strength by offering a more comprehensive framework with an overall accuracy of 94.6, effectively capturing the complex interdependencies among the variables. These findings highlight the strengths of both approaches: while the ANN model excels in analyzing individual variables, the ODE model offers a broader perspective by integrating all variables into a unified framework. This study contributes to developing economic forecasting models and provides valuable insights for decision-makers, particularly in times of economic uncertainty.

 This study explores the relationships between the USD opening exchange rate, the annual change rate of the Consumer Price Index CPI, the housing loan interest rate in Turkish lira, and the residential construction cost index from January 2015 to May 2024 using data from the Turkish Statistical Institute TUIK. Artificial Neural Networks ANN and Ordinary Differential Equations ODE methods were employed to model the interactions among these four variables. In the ANN approach, each variable was modeled as the dependent variable in turn, with the remaining three serving as independent variables, resulting in four distinct analyses. The ODE model, on the other hand, provided a holistic analysis by capturing the time-dependent relationships among all four variables simultaneously. The ANN model predictions achieved accuracy rates of 87.2 for the USD opening exchange rate, 91.4 for the CPI annual change rate, 85.9 for the housing loan interest rate, and 93.1 for the construction cost index. Meanwhile, the ODE model demonstrated its strength by offering a more comprehensive framework with an overall accuracy of 94.6, effectively capturing the complex interdependencies among the variables. These findings highlight the strengths of both approaches: while the ANN model excels in analyzing individual variables, the ODE model offers a broader perspective by integrating all variables into a unified framework. This study contributes to developing economic forecasting models and provides valuable insights for decision-makers, particularly in times of economic uncertainty.

Makale detayları

Comparative Forecasting of Some Key Economic Indicators Using Artificial Neural Networks and Ordinary Differential Equations: A Case Study of the Turkish Economy

Yazarlar:

Bahatdin Daşbaşı | Murat Taşyürek

Yayıncı:

Çanakkale Onsekiz Mart University

Ana Başlıklar:

Information Systems Development Methodologies and Practice

Dergi:

Journal of Advanced Research in Natural and Applied Sciences Vol.10 Issue 4, pp.922-937

Belge Tipi:

Makale

Yayın Tarihi:

12/2024

Belge Dili:

İngilizce

Document ID (Org No.):

2001338207

pISSN:

2757-5195

Turkish economy

artificial neural networks

ordinary differential equations

comparative forecasting

Sayı

Makale detayları

Comparative Forecasting of Some Key Economic Indicators Using Artificial Neural Networks and Ordinary Differential Equations: A Case Study of the Turkish Economy

Yazarlar: Bahatdin Daşbaşı | Murat Taşyürek

Yayıncı: Çanakkale Onsekiz Mart University

Ana Başlıklar: Information Systems Development Methodologies and Practice

Dergi: Journal of Advanced Research in Natural and Applied Sciences Vol.10 Issue 4, pp.922-937

Belge Tipi: Makale

Yayın Tarihi: 12/2024

Belge Dili: İngilizce

Document ID (Org No.): 2001338207

pISSN: 2757-5195

Turkish economy

artificial neural networks

ordinary differential equations

comparative forecasting

Sayı

Yazarlar:

Bahatdin Daşbaşı | Murat Taşyürek

Yayıncı:

Çanakkale Onsekiz Mart University

Ana Başlıklar:

Information Systems Development Methodologies and Practice

Dergi:

Journal of Advanced Research in Natural and Applied Sciences Vol.10 Issue 4, pp.922-937

Belge Tipi:

Makale

Yayın Tarihi:

12/2024

Belge Dili:

İngilizce

Document ID (Org No.):

2001338207

pISSN:

2757-5195