Ebru Yuksel Haliloglu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Ebru Yuksel Haliloglu has a total of 15 co-authors. Here are the top co-authors by number of shared publications:
- Hakan Berument
Department of Economics, Bilkent University, Ankara, Turkey
4 shared publications - Nukhet Dogan
Department of Econometrics, Ankara Haci Bayram Veli University, Ankara, Turkey
2 shared publications - Nurhan Hande Sevgi
Ufuk Üniversitesi
2 shared publications - Ahmet Burak Oğuz
Department of Emergency Medicine, Ankara University School of Medicine, Ankara, Turkey
1 shared publication - Aslı GÜNAY
ANKARA SOSYAL BİLİMLER ÜNİVERSİTESİ, SİYASAL BİLGİLER FAKÜLTESİ
1 shared publication - Ayça Koca
Department of Emergency Medicine, Ankara University School of Medicine, Ankara, Turkey
1 shared publication
Latest Publications
- Open Access
- 11 August 2026
- Published by Springer Nature in Neural Computing and Applications
- Vol. 38 (15) , 656
Enhancing LSTM performance for financial time series forecasting: the differencing approach - Diff-LSTM
Long Short-Term Memory (LSTM) networks are increasingly preferred over conventional econometric models for forecasting non-linear time series such as stock market data. This study proposes the “Diff-LSTM” approach, designed to enhance the forecasting capabilities of standard LSTM networks for non-stationary time series such as stock market data. Unlike conventional applications, the Diff-LSTM methodology trains the network on first-differenced – hence, stationary- series from which forecasts for the original series (in levels) are reconstructed by updating with the observed previous value of the level series. In LSTM networks, due to the limitations of activation functions such as sigmoid and tanh , a downward bias can be observed in forecasts of non-stationary series that have a unit root. Therefore, the proposed Diff-LSTM approach mitigates this downward bias, enabling more reliable forecasts. Diff-LSTM models are evaluated for forecasting accuracy across five major U.S. stock market indices using three cross-validation schemes and four performance measures. The current study demonstrates the superiority of the Diff-LSTM model, which outperformed standard LSTM models in 56 of 60 evaluated instances. This suggests that using Diff-LSTM models improves the forecasting performance considerably.
- 11 August 2026
- Published by Springer Nature in Neural Computing and Applications
- Vol. 38 (15) , 656
Open AccessEnhancing LSTM performance for financial time series forecasting: the differencing approach - Diff-LSTM
Long Short-Term Memory (LSTM) networks are increasingly preferred over conventional econometric models for forecasting non-linear time series such as stock market data. This study proposes the “Diff-LSTM” approach, designed to enhance the forecasting capabilities of standard LSTM networks for non-stationary time series such as stock market data. Unlike conventional applications, the Diff-LSTM methodology trains the network on first-differenced – hence, stationary- series from which forecasts for the original series (in levels) are reconstructed by updating with the observed previous value of the level series. In LSTM networks, due to the limitations of activation functions such as sigmoid and tanh , a downward bias can be observed in forecasts of non-stationary series that have a unit root. Therefore, the proposed Diff-LSTM approach mitigates this downward bias, enabling more reliable forecasts. Diff-LSTM models are evaluated for forecasting accuracy across five major U.S. stock market indices using three cross-validation schemes and four performance measures. The current study demonstrates the superiority of the Diff-LSTM model, which outperformed standard LSTM models in 56 of 60 evaluated instances. This suggests that using Diff-LSTM models improves the forecasting performance considerably.
- 5 January 2026
- Published by Gazi University Journal of Science in Gazi University Journal of Science
- No. Advanced O
Forecasting Performance Comparison: LSTM Model vs. Dynamic and Static Forecasting of ARIMA and Random Walk Models - Evidence from S&P500
This paper re-evaluates the forecasting performance of Long Short-Term Memory (LSTM) models against traditional Autoregressive Integrated Moving Average (ARIMA) and Random Walk (RW) models using S&P 500 index data. The LSTM models in the literature, which utilize observed lagged values for their forecasting (static forecasting), are compared to traditional models that use recursive, self-generated predictions for forecasting (dynamic forecasting). To provide comparable statistics, we gather relevant statistics for the static forecasting of ARIMA and RW models. We repeat the exercise for three different cross-validation schemes. The empirical evidence presented here suggests that the static forecasting power of static ARIMA and RW models consistently matches or outperforms that of LSTMs. Moreover, we find that while traditional models remain robust across varying sample sizes, the performance of LSTMs decreases when the training datasets are reduced. The empirical evidence presented here suggests that the reported superiority of LSTM in financial time series forecasting might be due to the forecasting method that is employed rather than a genuine predictive advantage.
- 5 January 2026
- Published by Gazi University Journal of Science in Gazi University Journal of Science
- No. Advanced O
Forecasting Performance Comparison: LSTM Model vs. Dynamic and Static Forecasting of ARIMA and Random Walk Models - Evidence from S&P500
This paper re-evaluates the forecasting performance of Long Short-Term Memory (LSTM) models against traditional Autoregressive Integrated Moving Average (ARIMA) and Random Walk (RW) models using S&P 500 index data. The LSTM models in the literature, which utilize observed lagged values for their forecasting (static forecasting), are compared to traditional models that use recursive, self-generated predictions for forecasting (dynamic forecasting). To provide comparable statistics, we gather relevant statistics for the static forecasting of ARIMA and RW models. We repeat the exercise for three different cross-validation schemes. The empirical evidence presented here suggests that the static forecasting power of static ARIMA and RW models consistently matches or outperforms that of LSTMs. Moreover, we find that while traditional models remain robust across varying sample sizes, the performance of LSTMs decreases when the training datasets are reduced. The empirical evidence presented here suggests that the reported superiority of LSTM in financial time series forecasting might be due to the forecasting method that is employed rather than a genuine predictive advantage.
- 23 December 2022
- Published by Taylor & Francis in Energy Sources, Part B: Economics, Planning, and Policy
- Vol. 18 (1)
Sectoral electricity consumption modeling with D-vine quantile regression: The US electricity market case
Efficient electricity demand planning is crucial for energy market actors. However, it is difficult as a consequence of climate change. We aim at investigating how climate variables (heating and cooling degree days) may affect electricity demand. By examining electricity consumption in various US sectors, we explore this relationship using parametric and non-parametric D-vine quantile regression models that exploits the dependence between covariates and allows sequential covariate selection. The results are compared against the classical linear quantile regression. We find a positive effect of the climatic variables on electricity consumption that is as heating and cooling degree days increase electricity demand rises in all sectors, and cooling need has a greater impact than heating need. Evidence suggests that residential and commercial electricity consumptions are affected the most, while industrial and transport sector consumptions are less sensitive. The D-vine quantile regression performs better than the linear quantile regression for almost all sectors.
- 23 December 2022
- Published by Taylor & Francis in Energy Sources, Part B: Economics, Planning, and Policy
- Vol. 18 (1)
Sectoral electricity consumption modeling with D-vine quantile regression: The US electricity market case
Efficient electricity demand planning is crucial for energy market actors. However, it is difficult as a consequence of climate change. We aim at investigating how climate variables (heating and cooling degree days) may affect electricity demand. By examining electricity consumption in various US sectors, we explore this relationship using parametric and non-parametric D-vine quantile regression models that exploits the dependence between covariates and allows sequential covariate selection. The results are compared against the classical linear quantile regression. We find a positive effect of the climatic variables on electricity consumption that is as heating and cooling degree days increase electricity demand rises in all sectors, and cooling need has a greater impact than heating need. Evidence suggests that residential and commercial electricity consumptions are affected the most, while industrial and transport sector consumptions are less sensitive. The D-vine quantile regression performs better than the linear quantile regression for almost all sectors.
- 5 October 2021
- Published by Elsevier in Economic Analysis and Policy
- Vol. 72, 492-505
Brent–Dubai oil spread: Basic drivers
- 5 October 2021
- Published by Elsevier in Economic Analysis and Policy
- Vol. 72, 492-505
Brent–Dubai oil spread: Basic drivers
- 29 April 2021
- Published by Galenos Yayinevi in Journal of Ankara University Faculty of Medicine
- Vol. 74 (1) , 43-51
Direct and Indirect Cost Analysis for Emergency Department Patients
- 29 April 2021
- Published by Galenos Yayinevi in Journal of Ankara University Faculty of Medicine
- Vol. 74 (1) , 43-51
Direct and Indirect Cost Analysis for Emergency Department Patients
- 16 April 2021
- Published by SAGE Publications in Global Business Review
- Vol. 25 (4) , 888-912
The Asymmetric Effects of Crude Oil Prices and Exchange Rates on Diesel Prices for 27 European Countries
Many studies have examined the asymmetric effect of US dollar-denominated crude oil prices on petroleum product prices. The ‘rockets and feathers’ argument suggests that a crude price increase raises petroleum product prices more than a corresponding decrease in crude prices lowers product prices. However, for the countries that do not use the US dollar as a medium of exchange, petroleum product prices are also affected by the exchange rates. This paper analysed the asymmetric effects of both US dollar-denominated crude oil prices and exchange rates on local currency-denominated diesel prices for 27 European countries in the short run as well as long run. The overall empirical evidence suggests that, in the short run, diesel prices react more to crude oil price increases than to a decrease, parallel to the ‘rockets and feathers’ argument. However, contrary to that argument, the long-run adjustment is the opposite. As for exchange rate shocks, again the ‘rockets and feathers’ argument holds and diesel prices respond more to exchange rate depreciation than appreciation in the short and long run.
- 16 April 2021
- Published by SAGE Publications in Global Business Review
- Vol. 25 (4) , 888-912
The Asymmetric Effects of Crude Oil Prices and Exchange Rates on Diesel Prices for 27 European Countries
Many studies have examined the asymmetric effect of US dollar-denominated crude oil prices on petroleum product prices. The ‘rockets and feathers’ argument suggests that a crude price increase raises petroleum product prices more than a corresponding decrease in crude prices lowers product prices. However, for the countries that do not use the US dollar as a medium of exchange, petroleum product prices are also affected by the exchange rates. This paper analysed the asymmetric effects of both US dollar-denominated crude oil prices and exchange rates on local currency-denominated diesel prices for 27 European countries in the short run as well as long run. The overall empirical evidence suggests that, in the short run, diesel prices react more to crude oil price increases than to a decrease, parallel to the ‘rockets and feathers’ argument. However, contrary to that argument, the long-run adjustment is the opposite. As for exchange rate shocks, again the ‘rockets and feathers’ argument holds and diesel prices respond more to exchange rate depreciation than appreciation in the short and long run.
- 1 August 2018
- Published by Universite Arastirmalari Dergisi in Üniversite Araştırmaları Dergisi
- Vol. 1 (2) , 36-42
A Case Study on Measuring Research Efficiency of Public Universities in Turkey
This paper measures the research efficiency and productivity of public universities founded before 1981 in Turkey over the period 2013-2016. Data Envelopment Analysis is applied to assess the relative research efficiency of these universities, while Malmquist Total Factor Productivity Index is used to measure the total factor productivity change with respect to research inputs of universities. The analysis made here differ from the similar studies in terms of the diversity of data set and the measurement of university research efficiency and productivity changes together. The results indicate that the number of relatively efficient universities according to their reserach inputs declines continuously over the years. Also, the productivity of research activities decreases except the period 2013-2014. It is additionally observed that 2.3% fall in research productivity of the universities is due to deterioration in both technological and technical efficiency over the yaers.
- 1 August 2018
- Published by Universite Arastirmalari Dergisi in Üniversite Araştırmaları Dergisi
- Vol. 1 (2) , 36-42
A Case Study on Measuring Research Efficiency of Public Universities in Turkey
This paper measures the research efficiency and productivity of public universities founded before 1981 in Turkey over the period 2013-2016. Data Envelopment Analysis is applied to assess the relative research efficiency of these universities, while Malmquist Total Factor Productivity Index is used to measure the total factor productivity change with respect to research inputs of universities. The analysis made here differ from the similar studies in terms of the diversity of data set and the measurement of university research efficiency and productivity changes together. The results indicate that the number of relatively efficient universities according to their reserach inputs declines continuously over the years. Also, the productivity of research activities decreases except the period 2013-2014. It is additionally observed that 2.3% fall in research productivity of the universities is due to deterioration in both technological and technical efficiency over the yaers.