Browsing by Author "Ünlü, Kamil Demirberk"

Now showing 1 - 17 of 17

Citation Count: 14
Daily PM₁₀, periodicity and harmonic regression model: The case of London
(Pergamon-elsevier Science Ltd, 2020) Okkaoglu, Yasin; Akdi, Yilmaz; Unlu, Kamil Demirberk; Industrial Engineering
One of the most important and distinguishable features of the climate driven data can be shown as the seasonality. Due to its nature air pollution data may have hourly, daily, weekly, monthly or even seasonal cycles. Many techniques such as non-linear time series analysis, machine learning algorithms and deterministic models, have been used to deal with this non-linear structure. Although, these models can capture the seasonality they can't identify the periodicity. Periodicity is beyond the seasonality, it is the hidden pattern of the time series. In this study, it is aimed to investigate the periodicity of daily Particulate Matter (PM10) of London between the periods 2014 and 2018. PM10 is the particulate matter of which aerodynamic diameter is less than 10 mu m. Firstly, periodogram based unit root test is used to check the stationarity of the investigated data. Afterwards, hidden periodic structure of the data is revealed. It is found that, it has five different cycle periods as 7 days, 25 days, 6 months, a year and 15 months. Lastly, it is shown that harmonic regression performs better in forecasting monthly and daily averages of the data.
Citation Count: 29
A deep learning approach to model daily particular matter of Ankara: key features and forecasting
(Springer, 2022) Akbal, Y.; Unlu, K. D.; Industrial Engineering
In this study, three different goals are pursued. Firstly, it is aimed to model particulate matter (PM) of Ankara, the capital of Turkey, by utilizing hybrid deep learning methodology. To do this, five different methodologies are proposed in which four of them are hybrid methods. Three different evaluation criteria as coefficient of determination (R-2), mean absolute error (MAE) and mean squared error (MSE) are used to compare the proposed methods. In the test set, the hybrid model which consists of feed-forward neural networks, convolution neural network and long short-term neural networks has the best performance with R-2 of 0.81, MSE of 73.07 and MAE of 5.6. Secondly, PM levels are categorized to form a prediction challenge in accordance with the World Health Organization standards. The particulate matter level is divided into two categories as being low or not, being moderate or not and being dangerous or not, it is shown that the proposed hybrid model which has the highest performance on forecasting, also worked perfectly in the classification task with accuracy of 94%. Finally, the effect of different pollutants and meteorological variables on the prediction of PM is investigated by employing ensemble machine learning methodology of random forest regression, extra tree regression and multiple linear regression. According to the results of the analysis, it is shown that the most important predictor variables of PM are its own lagged values, other pollutants, earth skin temperature and the wind speed.
Citation Count: 0
Determining Harmonic Fluctuations in Food Inflation
(World Scientific Publishing Co., 2022) Akdi,Y.; Ünlü,K.D.; Baş,C.; Karamanoğlu,Y.E.; Industrial Engineering
In this study, we start with a brief expression of consumer price index of Turkey. In the next step, we give the theoretical essentials of periodogram-based unit root and harmonic regression model. Periodogram-based unit root test is used to identify both the stationarity of data and periodicities. Periodicity is beyond seasonality; it is the hidden cycles in the data. Thus, it is harder to detect them compared to seasonal cycles. Harmonic-regression-type trigonometric regression models are useful in modeling data which have hidden periodicity. Afterward, the stationarity properties of monthly inflation and monthly food inflation of Turkey for the period between 2004 and 2020 are investigated. Standard augmented Dickey-Fuller unit root test shows that both series are integrated of order one. However, the periodogram-based unit root test shows that monthly inflation has unit root but monthly food inflation does not. After examining the unit root, the hidden cycles in the food inflation are revealed. The cycles in food inflation are important because they may trigger a headline inflation. The main contribution of this study is the identification of the hidden cycles in food inflation. It has cycles of approximately two, four, six and eight years. These cycles, in short, correspond to cycles of two years of consecutive periods. © 2022 by World Scientific Publishing Europe Ltd.
Citation Count: 8
Forecasting Air Quality in Tripoli: An Evaluation of Deep Learning Models for Hourly PM_2.5 Surface Mass Concentrations
(Mdpi, 2023) Esager, Marwa Winis Misbah; Unlu, Kamil Demirberk; Industrial Engineering
In this article, we aimed to study the forecasting of hourly PM2.5 surface mass concentrations in the city of Tripoli, Libya. We employed three state-of-the-art deep learning models, namely long short-term memory, gated recurrent unit, and convolutional neural networks, to forecast PM2.5 levels using univariate time series methodology. Our results revealed that the convolutional neural networks model performed the best, with a coefficient of variation of 99% and a mean absolute percentage error of 0.04. These findings provide valuable insights into the use of deep learning models for forecasting PM2.5 and can inform decision-making regarding air quality management in the city of Tripoli.
Citation Count: 0
Forecasting Direction of BIST 100 Index: An Integrated Machine Learning Approach
(Springer Science and Business Media B.V., 2021) Ünlü,K.D.; Potas,N.; Yılmaz,M.; Civil Engineering; Industrial Engineering
In recent years trends in analyzing and forecasting financial time series moves from classical Box-Jenkins methodology to machine learning algorithms because of the non-linearity and non-stationary of the time series. In this study, we employed a machine learning algorithm called support vector machine to predict the daily price direction of BIST 100 index. In addition, we use random forest algorithm for feature selection and showed that by removing some features from the model, performance of the model increases. © 2021, The Author(s), under exclusive license to Springer Nature Switzerland AG.
Citation Count: 0
Forecasting the BIST 100 Index with Support Vector Machines
(World Scientific Publishing Co., 2022) Ünlü,K.D.; Potas,N.; Ylmaz,M.; Industrial Engineering
Recent literature shows that statistical learning algorithms are powerful for forecasting financial time series. In this study, we model and forecast the Borsa Istanbul 100 Index by employing the machine learning algorithm, support vector machine. The dataset contains the highest price, lowest price, closing price and volume of the index for the period between July 2020 and June 2021.We utilize three different kernels. The empirical findings show that linear kernel gives the best result with coefficient of determination of 0.91 and root mean square error of 0.0062. The second best is polynomial kernel, and it is followed by radial basis kernel. The study shows that statistical learning algorithms can be thought of as an alternative to classical time series methodology in forecasting financial time series. © 2022 by World Scientific Publishing Europe Ltd.
Citation Count: 0
A hybrid deep learning methodology for wind power forecasting based on attention
(Taylor & Francis inc, 2024) Akbal, Yildirim; Unlu, Kamil Demirberk; Industrial Engineering; Mathematics
Wind energy, as a sustainable energy source, poses challenges in terms of storage. Therefore, careful planning is crucial to utilize it efficiently. Deep learning algorithms are gaining popularity for analyzing complex time series data. However, as the "no free lunch" theorem suggests, the trade-off is: they need a lot of data to achieve the benefits. This even brings up a severe challenge for time series analysis, as the availability of historical data is often limited. This study aims to address this issue by proposing a novel shallow deep learning approach for wind power forecasting. The proposed model utilizes a fusion of transformers, convolutional and recurrent neural networks to efficiently handle several time series simultaneously. The empirical evidence demonstrates that the suggested innovative method exhibits exceptional forecasting performance, as indicated by a coefficient of determination (R2) of 0.99. When the forecasting horizon reaches 48, the model's performance declines significantly. However, when dealing with long ranges, utilizing the mean as a metric rather than individual point estimates would yield superior results. Even when forecasting up to 96 hrs in advance, obtaining an R2 value of 0.50 is considered a noteworthy accomplishment in the context of average forecasting.
Citation Count: 8
Identifying the cycles in COVID-19 infection: the case of Turkey
(Taylor & Francis Ltd, 2023) Akdi, Yilmaz; Karamanoglu, Yunus Emre; Unlu, Kamil Demirberk; Bas, Cem; Industrial Engineering
The new coronavirus disease, called COVID-19, has spread extremely quickly to more than 200 countries since its detection in December 2019 in China. COVID-19 marks the return of a very old and familiar enemy. Throughout human history, disasters such as earthquakes, volcanic eruptions and even wars have not caused more human losses than lethal diseases, which are caused by viruses, bacteria and parasites. The first COVID-19 case was detected in Turkey on 12 March 2020 and researchers have since then attempted to examine periodicity in the number of daily new cases. One of the most curious questions in the pandemic process that affects the whole world is whether there will be a second wave. Such questions can be answered by examining any periodicities in the series of daily cases. Periodic series are frequently seen in many disciplines. An important method based on harmonic regression is the focus of the study. The main aim of this study is to identify the hidden periodic structure of the daily infected cases. Infected case of Turkey is analyzed by using periodogram-based methodology. Our results revealed that there are 4, 5 and 62 days cycles in the daily new cases of Turkey.
Citation Count: 13
Modeling and forecasting of monthly PM_2.5 emission of Paris by periodogram-based time series methodology
(Springer, 2021) Akdi, Yilmaz; Golveren, Elif; Unlu, Kamil Demirberk; Yucel, Mustafa Eray; Industrial Engineering
In this study, monthly particulate matter (PM2.5) of Paris for the period between January 2000 and December 2019 is investigated by utilizing a periodogram-based time series methodology. The main contribution of the study is modeling the PM2.5 of Paris by extracting the information purely from the examined time series data, where proposed model implicitly captures the effects of other factors, as all their periodic and seasonal effects reside in the air pollution data. Periodicity can be defined as the patterns embedded in the data other than seasonality, and it is crucial to understand the underlying periodic dynamics of air pollutants to better fight pollution. The method we use successfully captures and accounts for the periodicities, which could otherwise be mixed with seasonality under an alternative methodology. Upon the unit root test based on periodograms, it is revealed that the investigated data has periodicities of 1 year and 20 years, so harmonic regression is utilized as an alternative to Box-Jenkins methodology. As the harmonic regression displayed a better performance both in and out-of-sample forecasts, it can be considered as a powerful alternative to model and forecast time series with a periodic structure.
Citation Count: 8
A new generalized δ-shock model and its application to 1-out-of-(m+1):G cold standby system
(Elsevier Sci Ltd, 2023) Eryilmaz, Serkan; Unlu, Kamil Demirberk; Industrial Engineering
According to the classical delta-shock model, the system failure occurs upon the occurrence of a new shock that arrives in a time length less than delta, a given positive value. In this paper, a new generalized version of the delta-shock model is introduced. Under the proposed model, the system fails if there are m shocks that arrive in a time length less than delta after a previous shock, m >= 1. The mean time to failure of the system is approximated for both discretely and continuously distributed intershock time distributions. The usefulness of the model is also shown to study 1-out-of-(m + 1):G cold standby system. Illustrative numerical results are presented for geometric, exponential, discrete and continuous phase-type intershock time distributions.
Citation Count: 15
Periodicity in precipitation and temperature for monthly data of Turkey
(Springer Wien, 2021) Akdi, Yilmaz; Unlu, Kamil Demirberk; Industrial Engineering
In this study, we model and forecast monthly average temperature and monthly average precipitation of Turkey by employing periodogram-based time series methodology. We compare autoregressive integrated moving average methodology and harmonic regression. We show that harmonic regression performs better than the classical methodology in both time series. Also, we find that the monthly average temperature and monthly average precipitation have two different periodic structures of 6 months and 12 months which coincide with the seasonal pattern of the time series.
Citation Count: 0
Relationships Between Stock Markets: Causality Between G8 Countries and Turkey
(Springer international Publishing Ag, 2020) Unlu, Kamil Demirberk; Potas, Nihan; Yilmaz, Mehmet; Industrial Engineering
This study investigated relationships between stock markets in the Group of Eight (G8) countries (Canada, France, Germany, Italy, Japan, Russia, the UK, and the USA) and the Istanbul Stock Exchange (ISE) by estimating eight different vector autoregressions (VARs). We applied the Johansen and Juselius cointegration test to identify the long-run relations between the indices. The modified Granger causality test proposed by Toda and Yamamoto was conducted to identify the causality, then forecast variance decomposition and impulse response analysis were employed to explore the impacts of unexpected shocks in the G8 countries' stock markets on the ISE. The results showed that there was no cointegration between the ISE and the G8 countries' markets, but they still affected the ISE to different degrees, and the DAX-ISE 100, CAC 40-ISE 100, and FTSE MIB-ISE 100 causal relationships were bidirectional.
Citation Count: 2
Strategic Electricity Production Planning of Turkey via Mixed Integer Programming Based on Time Series Forecasting
(Mdpi, 2023) Yoruk, Gokay; Bac, Ugur; Yerlikaya-Ozkurt, Fatma; Unlu, Kamil Demirberk; Industrial Engineering
This study examines Turkey's energy planning in terms of strategic planning, energy policy, electricity production planning, technology selection, and environmental policies. A mixed integer optimization model is proposed for strategic electricity planning in Turkey. A set of energy resources is considered simultaneously in this research, and in addition to cost minimization, different strategic level policies, such as CO2 emission reduction policies, energy resource import/export restriction policies, and renewable energy promotion policies, are also considered. To forecast electricity demand over the planning horizon, a variety of forecasting techniques, including regression methods, exponential smoothing, Winter's method, and Autoregressive Integrated Moving Average methods, are used, and the best method is chosen using various error measures. The optimization model constructed for Turkey's Strategic Electricity Planning is obtained for two different planning intervals. The findings indicate that the use of renewable energy generation options, such as solar, wind, and hydroelectric alternatives, will increase significantly, while the use of fossil fuels in energy generation will decrease sharply. The findings of this study suggest a gradual increase in investments in renewable energy-based electricity production strategies are required to eventually replace fossil fuel alternatives. This change not only reduces investment, operation, and maintenance costs, but also reduces emissions in the long term.
Citation Count: 0
Türkiye gıda enflasyonunun dönemselliklerinin alt gruplar dikkate alınarak incelenmesi
(2022) Ünlü, Kamil Demirberk; Baş, Cem; Akdi, Yılmaz; Karamanoğlu, Yunus Emre; Industrial Engineering
Bu makale, Ocak 2004 - Haziran 2020 dönemleri için Türkiye aylık gıda enflasyonu zaman serilerindeki gizli periyodiklikleri alt grupları ile ampirik olarak tanımlamaktadır. Dönemsellik, mevsimselliğin ötesinde olan zaman serilerinin gizli döngüleridir. Bu gizli döngülerle başa çıkmak için periodogram tabanlı zaman serisi analizi kullanılır. Ayrıca, gelecekteki enflasyon oranları harmonik regresyon ile tahmin edilmektedir. Bu çalışmanın sonuçları, Türkiye gıda enflasyonunun yaklaşık 2 yıllık döngülere sahip olduğunu ve bunun çoğunlukla gıda ve alkolsüz içecek enflasyonundan kaynaklandığını ortaya koymaktadır.
Citation Count: 0
TÜRKİYE'DEKİ TRAFİK KAZALARININ PERİYODİK YAPISININ ARAŞTIRILMASI
(2021) Akdi, Yılmaz; Karamanoğlu, Yunus Emre; Ünlü, Kamil Demirberk; Baş, Cem; Industrial Engineering
Bu çalışmada, Türkiye'de 2019 yılında meydana gelen günlük trafik kazaları verilerine zaman serisi analizi uygulanmıştır. Çalışmada kullanılan verilerin en önemli özelliği kolluk birimleri tarafından günlük olarak tutulan resmi trafik kazası kayıtları olmasıdır. Bu verilerle ilgili olarak en uygun zaman serisi modeli belirlenmiş ve trafik kazalarında periyodik bileşenlerin olup olmadığı incelenmiştir. Verilerin birinci dereceden entegre olduğu görülmektedir. Bu durumda serinin birinci dereceden farkı istatistiksel sonuç açısından alınmıştır. Serinin grafikleri incelendiğinde, serilerde olası periyodikliğin bulunabileceği varsayımı ile Akdi ve Dickey (1998) tarafından önerilen periodogram temelli birim kök testi ile serinin durağanlığı da test edilmiş ve serinin % 10 anlamlılık düzeyinde durağan olduğu görülmüştür. Elde edilen sonuçlara göre 2019 yılında günlük trafik kaza sayılarında 33, 36.5 ve 73 günlük dönemlerin önemli olduğu tespit edilmiştir. 73 günlük sürenin Ramazan Bayramı ile Kurban Bayramı arasındaki döneme denk geldiği (iki dini bayram arasında 70 günlük bir ara vardır) gösterilmiştir.
Citation Count: 7
A two-step machine learning approach to predict S&P 500 bubbles
(Taylor & Francis Ltd, 2021) Kabran, Fatma Basoglu; Unlu, Kamil Demirberk; Industrial Engineering
In this paper, we are interested in predicting the bubbles in the S&P 500 stock market with a two-step machine learning approach that employs a real-time bubble detection test and support vector machine (SVM). SVM as a nonparametric binary classification technique is already a widely used method in financial time series forecasting. In the literature, a bubble is often defined as a situation where the asset price exceeds its fundamental value. As one of the early warning signals, prediction of bubbles is vital for policymakers and regulators who are responsible to take preemptive measures against the future crises. Therefore, many attempts have been made to understand the main factors in bubble formation and to predict them in their earlier phases. Our analysis consists of two steps. The first step is to identify the bubbles in the S&P 500 index using a widely recognized right-tailed unit root test. Then, SVM is employed to predict the bubbles by macroeconomic indicators. Also, we compare SVM with different supervised learning algorithms by usingk-fold cross-validation. The experimental results show that the proposed approach with high predictive power could be a favourable alternative in bubble prediction.
Citation Count: 14
A univariate time series methodology based on sequence-to-sequence learning for short to midterm wind power production
(Pergamon-elsevier Science Ltd, 2022) Akbal, Yildirim; Unlu, Kamil Demirberk; Industrial Engineering; Mathematics
The biggest wind farm of Turkey is placed at Manisa which is located in the Aegean Region. Electricity is a nonstorable commodity for that reason, it is very important to have a strong forecast and model of the potential electricity production to plan the electricity loads. In this study, the aim is to model and forecast electricity production of the wind farms located at Manisa by using a univariate model based on sequence-to-sequence learning. The forecasting range of the study is from short term to midterm. The strength of the proposed model is that; it only needs its own lagged value to make forecasts. The empirical evidences show that the model has high coefficient of variation (R-2) in short term and moderate R-2 in the midterm forecast. Although in the midrange forecasts R-2 slightly decreases mean squared error and mean absolute error shows that the model is accurate also in the midterm forecasts. The proposed model is not only strong in hourly electricity production forecasts but with a slight modification also in forecasting the minimum, maximum and average electricity production for a fixed range. This study concludes with two fresh and intriguing future research ideas.