Forecasting inflation accurately in a data-rich environment is a challenging task and an active r... more Forecasting inflation accurately in a data-rich environment is a challenging task and an active research field which still contains various unanswered methodological questions. One of them is how to find and extract the information with the most predictive power for a variable of interest when there are many highly correlated predictors, as in the inflation forecasting problem. Traditionally, factor models have been used to tackle this problem. However, a few recent studies have revealed that machine learning (ML) models such as random forests may offer some valuable solutions to the problem. This study encourages greater use of ML models with or without factor models by replacing the functional form of the forecast equation in a factor model with ML models or directly employing them with several feature selection techniques. This study adds new tree-based models to the analysis in the light of the recent findings in the literature. Moreover, it proposes the integration of feature selection techniques with Shapley values to find out concise explanations of the inflation predictions. The results obtained by a comprehensive set of experiments in an emerging country, Turkey, facing a high degree of volatility and uncertainty, indicate that tree-based ensemble models can be advantageous by providing better accuracy together with explainable predictions.
It is a well-established fact that energy consumption and production, as the primary sources of g... more It is a well-established fact that energy consumption and production, as the primary sources of greenhouse gases, contribute to climate change and global warming issues. The analysis and estimation of the factors that contribute to these harmful gases will be of great assistance in the development of policies to reduce carbon dioxide emissions. In addition to identifying the factors related to energy consumption and CO2 emissions, forecasting the variable of interest as accurately as possible has a key role in increasing the efficiency of energy strategies to be implemented. Unlike studies in the literature, this study not only forecasts the future value of energy consumption and CO2 emissions but also determines the relationship between the predictions and the influential variables by revealing the contribution of each variable to the prediction. For this purpose, the study proposes an interpretable forecasting framework based on values of the Shapley additive explanation (SHAP) to provide a simpler explanation of machine learning (ML) models in forecasting energy consumption and CO2 emissions. The results obtained show that the total electricity generation from different energy sources is found to be the most important variable interacting positively with both energy consumption and CO2 emissions. Also, the influence of the predictors on projections made before and after COVID-19 has changed dramatically. The proposed method may assist policymakers in making future energy investments and establishing energy laws more accurately and efficiently as it explains the drivers of the forecasts.
Gunumuzde yapilan para talebinin modellenmesine iliskin literaturu inceledigimizde, modelleme yon... more Gunumuzde yapilan para talebinin modellenmesine iliskin literaturu inceledigimizde, modelleme yontemi olarak cogunlukla kointegrasyon testlerinin kullanilmis oldugu ve bu testlerden Engle-Granger (1987) yani sira Johansen-Juselius (1990) tarafindan onerilen kointegrasyon tekniklerinin bircok arastirmaci tarafindan cogu ulkenin para talebinin modellenmesinde uygulandigi gorulmektedir. Para politikasinin uygulanmasi asamasinda politik tutumlarin degerlendirilmesinin yani sira para politikasi kapsaminda alinacak kararlar, para ve diger makroekonomik degiskenler arasindaki iliskinin incelenmesine bagli oldugundan, farkli fonksiyonel iliskileri degerlendirmemize izin veren farkli modelleme yontemlerinin ele alinarak degerlendirilmesi onem kazanmaktadir. Bu calismanin amaci, literaturde yaygin bir bicimde kullanilan ekonometrik modelleme yontemine karsilik alternatif bir yontem ortaya koymaktir. Bu nedenle, calismada kointegrasyon yontemlerinden biri olan Engle-Granger (1987) tarafindan o...
As known; goal programming, which is a type of multi-objective programming that its priority is s... more As known; goal programming, which is a type of multi-objective programming that its priority is satisfaction and based on optimization, allows decision maker to design an optimal system. Hence; a solution set provides simultaneous satisfactory is determined. Fuzzy goal programming model is analyzed in two different forms with respect to objective preemptive priority structure. In the first form, an importance level of preferences is specified quantitatively. In the second form, preference priority is added to model as linguistic information. In fuzzy goal programming; additive model which is adapted to mentioned two forms is developed by Tiwari, Dharmar and Rao (1987). Goal programming based on more satisfaction instead of optimization. Thus, using of fuzzy logic in this area is appealing. Goal programming approaches based on fuzzy logic is developed by the time of progress and every approach is aimed at getting better models. A model that allows preemptive priority in fuzzy goal pr...
Neural networks are one of the widely-used time series forecasting methods in time series applica... more Neural networks are one of the widely-used time series forecasting methods in time series applications. Among different neural network architectures and learning algorithms, the most popular choice is the feedforward Multilayer Perceptron (MLP). However, it suffers from some drawbacks such as getting trapped in local minima, human intervention during the stage of training, and limitations in architecture design. The aims of this study were twofold. The first was to employ NeuroEvolution of Augmenting Topologies (NEAT), which has many successful applications in numerous fields. In this paper, we applied it to time series forecasting for the first time and compared its performance with that of the MLP. The second aim was to analyse the performance resulting from the pairwise combination of these methods. In general, the results suggested that the forecasts from the NEAT algorithm were more accurate than those of the MLP. The results also showed that pairwise combined forecasts in gene...
... Hakemli mi?: Evet. Yazar(lar): Akdeniz, Ahmet (Yazar), Aras, Serkan (Yazar),. Emeği Geçen(ler... more ... Hakemli mi?: Evet. Yazar(lar): Akdeniz, Ahmet (Yazar), Aras, Serkan (Yazar),. Emeği Geçen(ler): ... Sözü edilen iki yapıdan birincisi tercihlerin önem derecesinin sayısal olarak belirlendiği sistem; ikincisi ise tercih önceliğinin sözel bilgi olarak modele katıldığı sistemsel yapıdır. ...
Günümüzde yapılan para talebinin modellenmesine ilişkin literatürü incelediğimizde, modelleme yön... more Günümüzde yapılan para talebinin modellenmesine ilişkin literatürü incelediğimizde, modelleme yöntemi olarak çoğunlukla kointegrasyon testlerinin kullanılmış olduğu ve bu testlerden Engle-Granger 1987) yanı sıra Johansen-Juselius(1990) tarafından önerilen ...
When the literature regarding applications of neural networks is investigated, it appears that a ... more When the literature regarding applications of neural networks is investigated, it appears that a substantial issue is what size the training data should be when modelling a time series through neural networks. The aim of this paper is to determine the size of training data to be used to construct a forecasting model via a multiple-breakpoint test and compare its performance with two general methods, namely, using all available data and using just two years of data. Furthermore, the importance of the selection of the final neural network model is investigated in detail. The results obtained from daily crude oil prices indicate that the data from the last structural change lead to simpler architectures of neural networks and have an advantage in reaching more accurate forecasts in terms of MAE value. In addition, the statistical tests show that there is a statistically significant interaction between data size and stopping rule.
Machine learning techniques have been used frequently for volatility forecasting. However, previo... more Machine learning techniques have been used frequently for volatility forecasting. However, previous studies have built these hybrid models in a form of a first-order GARCH(1,1) process by following general use for GARCH models. But the way of estimating parameters for GARCH and machine learning models differs considerably. Hence, we have investigated the effect of different model orders of the GARCH process on the volatility forecasts of Bitcoin obtained by the four most used machine learning models. Furthermore, we have proposed a stacking ensemble methodology based on GARCH hybrid models to improve the results further. The proposed stacking ensemble methodology utilizes the techniques of feature selection and feature extraction to reduce the dimension of the predictors before meta-learning. The results show that using higher model orders increases the accuracy of volatility forecasts for hybrid GARCH models. Also, the proposed stacking ensemble with LASSO produces forecasts superior to almost all hybrid models and better than the ordinary stacking ensemble.
KRİPTO PARA FİYATLARININ KLASİK VE YAPAY SİNİR AĞI MODELLERİ İLE TAHMİNİ, 2019
Günümüzde kripto para birimlerinin
önemi gittikçe artmaktadır. Kripto para birimleri
sanal oyun p... more Günümüzde kripto para birimlerinin önemi gittikçe artmaktadır. Kripto para birimleri sanal oyun platformlarında kullanılırken, şu an pek çok kurum ve kuruluş tarafından ödeme aracı olarak kullanılmaktadır. Güvenlik risklerine karşı blockchain (Blok Zinciri) adı verilen algoritması ile üretimi sağlanmaktadır. Kripto para fiyatlarının doğru olarak tahmin edilmesi yatırımcı ve karar vericiler açısından büyük önem taşımaktadır. Bu çalışma kapsamında en çok kullanılan dört kripto para birimine (Bitcoin, Ethereum, Ripple, Litecoin) ait fiyat değerleri tahmin edilmiştir. Çoklu kırılma testinden yararlanılarak her seriye ait kırılmalar belirlenerek analiz genişletilmiştir. Ele alınan sanal para değerlerini doğru bir şekilde tahmin etmek amacıyla hem klasik zaman serisi modellerinden hem de üç farklı tür yapay sinir ağı modelinden faydalanılmıştır. Ayrıca elde edilen tahminler üzerinde basit birleştirilme teknikleri uygulanmıştır. Rassal yürüyüşün egemen olduğu bu seriler arasından, özellikle işlem hacmi ve bilinilirliği en fazla olan Bitcoin sanal parasında rassal yürüyüş modelinden daha iyi sonuçlar elde edildiği gözlemlenmiştir.
Comparing and Combining MLP and NEAT for Time Series Forecasting, 2017
Neural networks are one of the widely-used time series forecasting methods in time series applica... more Neural networks are one of the widely-used time series forecasting methods in time series applications. Among different neural network architectures and learning algorithms, the most popular choice is the feedforward Multilayer Perceptron (MLP). However, it suffers from some drawbacks such as getting trapped in local minima, human intervention during the stage of training, and limitations in architecture design. The aims of this study were twofold. The first was to employ NeuroEvolution of Augmenting Topologies (NEAT), which has many successful applications in numerous fields. In this paper, we applied it to time series forecasting for the first time and compared its performance with that of the MLP. The second aim was to analyse the performance resulting from the pairwise combination of these methods. In general, the results suggested that the forecasts from the NEAT algorithm were more accurate than those of the MLP. The results also showed that pairwise combined forecasts in general were better than single forecasts. The best forecasts of all were obtained by pairwise combination of MLP and NEAT.
Data Size Requirement for Forecasting Daily Crude Oil Price with Neural Networks, 2019
When the literature regarding applications of neural networks is investigated, it appears that a ... more When the literature regarding applications of neural networks is investigated, it appears that a substantial issue is what size the training data should be when modelling a time series through neural networks. The aim of this paper is to determine the size of training data to be used to construct a forecasting model via a multiple-breakpoint test and compare its performance with two general methods, namely, using all available data and using just two years of data. Furthermore, the importance of the selection of the final neural network model is investigated in detail. The results obtained from daily crude oil prices indicate that the data from the last structural change lead to simpler architectures of neural networks and have an advantage in reaching more accurate forecasts in terms of MAE value. In addition, the statistical tests show that there is a statistically significant interaction between data size and stopping rule.
To improve the forecasting accuracies, researchers have long been using various combination
techn... more To improve the forecasting accuracies, researchers have long been using various combination techniques. In particular, the use of dissimilar methods for forecasting time series data is expected to provide superior results. Although numerous combination techniques have been proposed until date, the simple combination techniques —such as mean and median —maintain their strength, popularity, and utility. This paper proposes a new combination method based on the mean and median combination methods so as to combine the advantages of both these methods. The proposed combination technique attempts to utilize the strong aspects of each method and minimize the risk that arises from the selection of the combination method with poor performance. In order to depict the potential power of the proposed combining method, well-known six real-world time series data were used. Our results indicate that the proposed method presents with promising performances. In addition, a nonparametric statistical test was exploited to reveal the superiority of the proposed method over the single methods and other forecast combination methods from all of the investigated data sets.
Journal of Business Economics and Management, 2017
In today's competitive global economy, businesses must adjust themselves constantly to ever-chang... more In today's competitive global economy, businesses must adjust themselves constantly to ever-changing markets. Therefore, predicting future events in the marketplace is crucial to the maintenance of successful business activities. In this study, sales forecasts for a global furniture retailer operating in Turkey were made using state space models, ARIMA and ARFIMA models, neural networks, and Adaptive Network-based Fuzzy Inference System (ANFIS). Also, the forecasting performances of some widely used combining methods were evaluated by comparison with the weekly sales data for ten products. According to the best of our knowledge, this study is the first time that the recently developed state space models, also called ETS (Error-Trend-Seasonal) models, and the ANFIS model have been tested within combining methods for forecasting retail sales. Analysis of the results of the single models in isolation indicated that none of them outperformed all the others across all the time series investigated. However, the empirical results suggested that most of the combined forecasts examined could achieve statistically significant increases in forecasting accuracy compared with individual models and with the forecasts generated by the company's current system.
Although artificial neural networks have recently gained importance in time series applications, ... more Although artificial neural networks have recently gained importance in time series applications, some methodological shortcomings still continue to exist. One of these shortcomings is the selection of the final neural network model to be used to evaluate its performance in test set among many neural networks. The general way to overcome this problem is to divide data sets into training, validation, and test sets and also to select a neural network model that provides the smallest error value in the validation set. However, it is likely that the selected neural network model would be overfitting the validation data. This paper proposes a new model selection strategy (IHTS) for forecasting with neural networks. The proposed selection strategy first determines the numbers of input and hidden units, and then, selects a neural network model from various trials caused by different initial weights by considering validation and training performances of each neural network model. It is observed that the proposed selection strategy improves the performance of the neural networks statistically as compared with the classic model selection method in the simulated and real data sets. Also, it exhibits some robustness against the size of the validation data.
Forecasting inflation accurately in a data-rich environment is a challenging task and an active r... more Forecasting inflation accurately in a data-rich environment is a challenging task and an active research field which still contains various unanswered methodological questions. One of them is how to find and extract the information with the most predictive power for a variable of interest when there are many highly correlated predictors, as in the inflation forecasting problem. Traditionally, factor models have been used to tackle this problem. However, a few recent studies have revealed that machine learning (ML) models such as random forests may offer some valuable solutions to the problem. This study encourages greater use of ML models with or without factor models by replacing the functional form of the forecast equation in a factor model with ML models or directly employing them with several feature selection techniques. This study adds new tree-based models to the analysis in the light of the recent findings in the literature. Moreover, it proposes the integration of feature selection techniques with Shapley values to find out concise explanations of the inflation predictions. The results obtained by a comprehensive set of experiments in an emerging country, Turkey, facing a high degree of volatility and uncertainty, indicate that tree-based ensemble models can be advantageous by providing better accuracy together with explainable predictions.
It is a well-established fact that energy consumption and production, as the primary sources of g... more It is a well-established fact that energy consumption and production, as the primary sources of greenhouse gases, contribute to climate change and global warming issues. The analysis and estimation of the factors that contribute to these harmful gases will be of great assistance in the development of policies to reduce carbon dioxide emissions. In addition to identifying the factors related to energy consumption and CO2 emissions, forecasting the variable of interest as accurately as possible has a key role in increasing the efficiency of energy strategies to be implemented. Unlike studies in the literature, this study not only forecasts the future value of energy consumption and CO2 emissions but also determines the relationship between the predictions and the influential variables by revealing the contribution of each variable to the prediction. For this purpose, the study proposes an interpretable forecasting framework based on values of the Shapley additive explanation (SHAP) to provide a simpler explanation of machine learning (ML) models in forecasting energy consumption and CO2 emissions. The results obtained show that the total electricity generation from different energy sources is found to be the most important variable interacting positively with both energy consumption and CO2 emissions. Also, the influence of the predictors on projections made before and after COVID-19 has changed dramatically. The proposed method may assist policymakers in making future energy investments and establishing energy laws more accurately and efficiently as it explains the drivers of the forecasts.
Gunumuzde yapilan para talebinin modellenmesine iliskin literaturu inceledigimizde, modelleme yon... more Gunumuzde yapilan para talebinin modellenmesine iliskin literaturu inceledigimizde, modelleme yontemi olarak cogunlukla kointegrasyon testlerinin kullanilmis oldugu ve bu testlerden Engle-Granger (1987) yani sira Johansen-Juselius (1990) tarafindan onerilen kointegrasyon tekniklerinin bircok arastirmaci tarafindan cogu ulkenin para talebinin modellenmesinde uygulandigi gorulmektedir. Para politikasinin uygulanmasi asamasinda politik tutumlarin degerlendirilmesinin yani sira para politikasi kapsaminda alinacak kararlar, para ve diger makroekonomik degiskenler arasindaki iliskinin incelenmesine bagli oldugundan, farkli fonksiyonel iliskileri degerlendirmemize izin veren farkli modelleme yontemlerinin ele alinarak degerlendirilmesi onem kazanmaktadir. Bu calismanin amaci, literaturde yaygin bir bicimde kullanilan ekonometrik modelleme yontemine karsilik alternatif bir yontem ortaya koymaktir. Bu nedenle, calismada kointegrasyon yontemlerinden biri olan Engle-Granger (1987) tarafindan o...
As known; goal programming, which is a type of multi-objective programming that its priority is s... more As known; goal programming, which is a type of multi-objective programming that its priority is satisfaction and based on optimization, allows decision maker to design an optimal system. Hence; a solution set provides simultaneous satisfactory is determined. Fuzzy goal programming model is analyzed in two different forms with respect to objective preemptive priority structure. In the first form, an importance level of preferences is specified quantitatively. In the second form, preference priority is added to model as linguistic information. In fuzzy goal programming; additive model which is adapted to mentioned two forms is developed by Tiwari, Dharmar and Rao (1987). Goal programming based on more satisfaction instead of optimization. Thus, using of fuzzy logic in this area is appealing. Goal programming approaches based on fuzzy logic is developed by the time of progress and every approach is aimed at getting better models. A model that allows preemptive priority in fuzzy goal pr...
Neural networks are one of the widely-used time series forecasting methods in time series applica... more Neural networks are one of the widely-used time series forecasting methods in time series applications. Among different neural network architectures and learning algorithms, the most popular choice is the feedforward Multilayer Perceptron (MLP). However, it suffers from some drawbacks such as getting trapped in local minima, human intervention during the stage of training, and limitations in architecture design. The aims of this study were twofold. The first was to employ NeuroEvolution of Augmenting Topologies (NEAT), which has many successful applications in numerous fields. In this paper, we applied it to time series forecasting for the first time and compared its performance with that of the MLP. The second aim was to analyse the performance resulting from the pairwise combination of these methods. In general, the results suggested that the forecasts from the NEAT algorithm were more accurate than those of the MLP. The results also showed that pairwise combined forecasts in gene...
... Hakemli mi?: Evet. Yazar(lar): Akdeniz, Ahmet (Yazar), Aras, Serkan (Yazar),. Emeği Geçen(ler... more ... Hakemli mi?: Evet. Yazar(lar): Akdeniz, Ahmet (Yazar), Aras, Serkan (Yazar),. Emeği Geçen(ler): ... Sözü edilen iki yapıdan birincisi tercihlerin önem derecesinin sayısal olarak belirlendiği sistem; ikincisi ise tercih önceliğinin sözel bilgi olarak modele katıldığı sistemsel yapıdır. ...
Günümüzde yapılan para talebinin modellenmesine ilişkin literatürü incelediğimizde, modelleme yön... more Günümüzde yapılan para talebinin modellenmesine ilişkin literatürü incelediğimizde, modelleme yöntemi olarak çoğunlukla kointegrasyon testlerinin kullanılmış olduğu ve bu testlerden Engle-Granger 1987) yanı sıra Johansen-Juselius(1990) tarafından önerilen ...
When the literature regarding applications of neural networks is investigated, it appears that a ... more When the literature regarding applications of neural networks is investigated, it appears that a substantial issue is what size the training data should be when modelling a time series through neural networks. The aim of this paper is to determine the size of training data to be used to construct a forecasting model via a multiple-breakpoint test and compare its performance with two general methods, namely, using all available data and using just two years of data. Furthermore, the importance of the selection of the final neural network model is investigated in detail. The results obtained from daily crude oil prices indicate that the data from the last structural change lead to simpler architectures of neural networks and have an advantage in reaching more accurate forecasts in terms of MAE value. In addition, the statistical tests show that there is a statistically significant interaction between data size and stopping rule.
Machine learning techniques have been used frequently for volatility forecasting. However, previo... more Machine learning techniques have been used frequently for volatility forecasting. However, previous studies have built these hybrid models in a form of a first-order GARCH(1,1) process by following general use for GARCH models. But the way of estimating parameters for GARCH and machine learning models differs considerably. Hence, we have investigated the effect of different model orders of the GARCH process on the volatility forecasts of Bitcoin obtained by the four most used machine learning models. Furthermore, we have proposed a stacking ensemble methodology based on GARCH hybrid models to improve the results further. The proposed stacking ensemble methodology utilizes the techniques of feature selection and feature extraction to reduce the dimension of the predictors before meta-learning. The results show that using higher model orders increases the accuracy of volatility forecasts for hybrid GARCH models. Also, the proposed stacking ensemble with LASSO produces forecasts superior to almost all hybrid models and better than the ordinary stacking ensemble.
KRİPTO PARA FİYATLARININ KLASİK VE YAPAY SİNİR AĞI MODELLERİ İLE TAHMİNİ, 2019
Günümüzde kripto para birimlerinin
önemi gittikçe artmaktadır. Kripto para birimleri
sanal oyun p... more Günümüzde kripto para birimlerinin önemi gittikçe artmaktadır. Kripto para birimleri sanal oyun platformlarında kullanılırken, şu an pek çok kurum ve kuruluş tarafından ödeme aracı olarak kullanılmaktadır. Güvenlik risklerine karşı blockchain (Blok Zinciri) adı verilen algoritması ile üretimi sağlanmaktadır. Kripto para fiyatlarının doğru olarak tahmin edilmesi yatırımcı ve karar vericiler açısından büyük önem taşımaktadır. Bu çalışma kapsamında en çok kullanılan dört kripto para birimine (Bitcoin, Ethereum, Ripple, Litecoin) ait fiyat değerleri tahmin edilmiştir. Çoklu kırılma testinden yararlanılarak her seriye ait kırılmalar belirlenerek analiz genişletilmiştir. Ele alınan sanal para değerlerini doğru bir şekilde tahmin etmek amacıyla hem klasik zaman serisi modellerinden hem de üç farklı tür yapay sinir ağı modelinden faydalanılmıştır. Ayrıca elde edilen tahminler üzerinde basit birleştirilme teknikleri uygulanmıştır. Rassal yürüyüşün egemen olduğu bu seriler arasından, özellikle işlem hacmi ve bilinilirliği en fazla olan Bitcoin sanal parasında rassal yürüyüş modelinden daha iyi sonuçlar elde edildiği gözlemlenmiştir.
Comparing and Combining MLP and NEAT for Time Series Forecasting, 2017
Neural networks are one of the widely-used time series forecasting methods in time series applica... more Neural networks are one of the widely-used time series forecasting methods in time series applications. Among different neural network architectures and learning algorithms, the most popular choice is the feedforward Multilayer Perceptron (MLP). However, it suffers from some drawbacks such as getting trapped in local minima, human intervention during the stage of training, and limitations in architecture design. The aims of this study were twofold. The first was to employ NeuroEvolution of Augmenting Topologies (NEAT), which has many successful applications in numerous fields. In this paper, we applied it to time series forecasting for the first time and compared its performance with that of the MLP. The second aim was to analyse the performance resulting from the pairwise combination of these methods. In general, the results suggested that the forecasts from the NEAT algorithm were more accurate than those of the MLP. The results also showed that pairwise combined forecasts in general were better than single forecasts. The best forecasts of all were obtained by pairwise combination of MLP and NEAT.
Data Size Requirement for Forecasting Daily Crude Oil Price with Neural Networks, 2019
When the literature regarding applications of neural networks is investigated, it appears that a ... more When the literature regarding applications of neural networks is investigated, it appears that a substantial issue is what size the training data should be when modelling a time series through neural networks. The aim of this paper is to determine the size of training data to be used to construct a forecasting model via a multiple-breakpoint test and compare its performance with two general methods, namely, using all available data and using just two years of data. Furthermore, the importance of the selection of the final neural network model is investigated in detail. The results obtained from daily crude oil prices indicate that the data from the last structural change lead to simpler architectures of neural networks and have an advantage in reaching more accurate forecasts in terms of MAE value. In addition, the statistical tests show that there is a statistically significant interaction between data size and stopping rule.
To improve the forecasting accuracies, researchers have long been using various combination
techn... more To improve the forecasting accuracies, researchers have long been using various combination techniques. In particular, the use of dissimilar methods for forecasting time series data is expected to provide superior results. Although numerous combination techniques have been proposed until date, the simple combination techniques —such as mean and median —maintain their strength, popularity, and utility. This paper proposes a new combination method based on the mean and median combination methods so as to combine the advantages of both these methods. The proposed combination technique attempts to utilize the strong aspects of each method and minimize the risk that arises from the selection of the combination method with poor performance. In order to depict the potential power of the proposed combining method, well-known six real-world time series data were used. Our results indicate that the proposed method presents with promising performances. In addition, a nonparametric statistical test was exploited to reveal the superiority of the proposed method over the single methods and other forecast combination methods from all of the investigated data sets.
Journal of Business Economics and Management, 2017
In today's competitive global economy, businesses must adjust themselves constantly to ever-chang... more In today's competitive global economy, businesses must adjust themselves constantly to ever-changing markets. Therefore, predicting future events in the marketplace is crucial to the maintenance of successful business activities. In this study, sales forecasts for a global furniture retailer operating in Turkey were made using state space models, ARIMA and ARFIMA models, neural networks, and Adaptive Network-based Fuzzy Inference System (ANFIS). Also, the forecasting performances of some widely used combining methods were evaluated by comparison with the weekly sales data for ten products. According to the best of our knowledge, this study is the first time that the recently developed state space models, also called ETS (Error-Trend-Seasonal) models, and the ANFIS model have been tested within combining methods for forecasting retail sales. Analysis of the results of the single models in isolation indicated that none of them outperformed all the others across all the time series investigated. However, the empirical results suggested that most of the combined forecasts examined could achieve statistically significant increases in forecasting accuracy compared with individual models and with the forecasts generated by the company's current system.
Although artificial neural networks have recently gained importance in time series applications, ... more Although artificial neural networks have recently gained importance in time series applications, some methodological shortcomings still continue to exist. One of these shortcomings is the selection of the final neural network model to be used to evaluate its performance in test set among many neural networks. The general way to overcome this problem is to divide data sets into training, validation, and test sets and also to select a neural network model that provides the smallest error value in the validation set. However, it is likely that the selected neural network model would be overfitting the validation data. This paper proposes a new model selection strategy (IHTS) for forecasting with neural networks. The proposed selection strategy first determines the numbers of input and hidden units, and then, selects a neural network model from various trials caused by different initial weights by considering validation and training performances of each neural network model. It is observed that the proposed selection strategy improves the performance of the neural networks statistically as compared with the classic model selection method in the simulated and real data sets. Also, it exhibits some robustness against the size of the validation data.
IDENTIFYING THE SUPERIOR GARCH MODELS THROUGH MODEL CONFIDENCE SET PROCEDURE, 2019
The modelling and forecasting of the volatility of the stock markets has been a topic of great in... more The modelling and forecasting of the volatility of the stock markets has been a topic of great interest in the world of finance.The motivation of the study was threefold: (1) we employ elev- en GARCH models such as, sGARCH, EGARCH, GJR-GARCH, APGARCH, IGARCH, csGARCH, FIGARCH, AVGARCH, TGARCH, NGARCH and NAGARCH to produce better forecasts, (2) we use ten different distributions, such as normal (norm), gener- alized error distribution (ged), student t-distribution (std), skewed normal distribution (snorm), skewed generalized error distribution (sged), skewed student t-distribution (sstd), generalized hyperbolic distribution (ghyp), normal inverse gaussian distribution (nig), gen- eralized hyperbolic student t-distribution (ghst) and Johnson’s Su distribution (jsu), and (3) we propose a process for the model and the distribution by using a fair selection platform which is named ‘model confidence set’.In this study, we aim to provide an empirical foundation for a better understanding of the ability of different GARCH models with different distributions in terms of out-of-sample forecasting performace.
Uploads
Papers by serkan aras
önemi gittikçe artmaktadır. Kripto para birimleri
sanal oyun platformlarında kullanılırken, şu an pek
çok kurum ve kuruluş tarafından ödeme aracı olarak
kullanılmaktadır. Güvenlik risklerine karşı
blockchain (Blok Zinciri) adı verilen algoritması ile
üretimi sağlanmaktadır. Kripto para fiyatlarının
doğru olarak tahmin edilmesi yatırımcı ve karar
vericiler açısından büyük önem taşımaktadır. Bu
çalışma kapsamında en çok kullanılan dört kripto
para birimine (Bitcoin, Ethereum, Ripple, Litecoin)
ait fiyat değerleri tahmin edilmiştir. Çoklu kırılma
testinden yararlanılarak her seriye ait kırılmalar
belirlenerek analiz genişletilmiştir. Ele alınan sanal
para değerlerini doğru bir şekilde tahmin etmek
amacıyla hem klasik zaman serisi modellerinden
hem de üç farklı tür yapay sinir ağı modelinden
faydalanılmıştır. Ayrıca elde edilen tahminler
üzerinde basit birleştirilme teknikleri uygulanmıştır.
Rassal yürüyüşün egemen olduğu bu seriler
arasından, özellikle işlem hacmi ve bilinilirliği en
fazla olan Bitcoin sanal parasında rassal yürüyüş
modelinden daha iyi sonuçlar elde edildiği
gözlemlenmiştir.
techniques. In particular, the use of dissimilar methods for forecasting time series data is expected to
provide superior results. Although numerous combination techniques have been proposed until date,
the simple combination techniques —such as mean and median —maintain their strength, popularity,
and utility. This paper proposes a new combination method based on the mean and median
combination methods so as to combine the advantages of both these methods. The proposed
combination technique attempts to utilize the strong aspects of each method and minimize the risk
that arises from the selection of the combination method with poor performance. In order to depict
the potential power of the proposed combining method, well-known six real-world time series data
were used. Our results indicate that the proposed method presents with promising performances. In
addition, a nonparametric statistical test was exploited to reveal the superiority of the proposed
method over the single methods and other forecast combination methods from all of the investigated
data sets.
önemi gittikçe artmaktadır. Kripto para birimleri
sanal oyun platformlarında kullanılırken, şu an pek
çok kurum ve kuruluş tarafından ödeme aracı olarak
kullanılmaktadır. Güvenlik risklerine karşı
blockchain (Blok Zinciri) adı verilen algoritması ile
üretimi sağlanmaktadır. Kripto para fiyatlarının
doğru olarak tahmin edilmesi yatırımcı ve karar
vericiler açısından büyük önem taşımaktadır. Bu
çalışma kapsamında en çok kullanılan dört kripto
para birimine (Bitcoin, Ethereum, Ripple, Litecoin)
ait fiyat değerleri tahmin edilmiştir. Çoklu kırılma
testinden yararlanılarak her seriye ait kırılmalar
belirlenerek analiz genişletilmiştir. Ele alınan sanal
para değerlerini doğru bir şekilde tahmin etmek
amacıyla hem klasik zaman serisi modellerinden
hem de üç farklı tür yapay sinir ağı modelinden
faydalanılmıştır. Ayrıca elde edilen tahminler
üzerinde basit birleştirilme teknikleri uygulanmıştır.
Rassal yürüyüşün egemen olduğu bu seriler
arasından, özellikle işlem hacmi ve bilinilirliği en
fazla olan Bitcoin sanal parasında rassal yürüyüş
modelinden daha iyi sonuçlar elde edildiği
gözlemlenmiştir.
techniques. In particular, the use of dissimilar methods for forecasting time series data is expected to
provide superior results. Although numerous combination techniques have been proposed until date,
the simple combination techniques —such as mean and median —maintain their strength, popularity,
and utility. This paper proposes a new combination method based on the mean and median
combination methods so as to combine the advantages of both these methods. The proposed
combination technique attempts to utilize the strong aspects of each method and minimize the risk
that arises from the selection of the combination method with poor performance. In order to depict
the potential power of the proposed combining method, well-known six real-world time series data
were used. Our results indicate that the proposed method presents with promising performances. In
addition, a nonparametric statistical test was exploited to reveal the superiority of the proposed
method over the single methods and other forecast combination methods from all of the investigated
data sets.