Performance Comparisons of Deep Learning and ARIMA:  A Borsa Istanbul Stock Example

Caner Erden

Makale detayları

Performance Comparisons of Deep Learning and ARIMA: A Borsa Istanbul Stock Example

Yazarlar:

Caner Erden

Yayıncı:

Manisa Celal Bayar University

Ana Başlıklar:

Mali Tablo Analizi

Dergi:

Yönetim ve Ekonomi Dergisi Vol.30 Issue 3, pp.419-438

Belge Tipi:

Makale

Yayın Tarihi:

9/2023

Belge Dili:

Türkçe

Document ID (Org No.):

2001298303

pISSN:

1302-0064

Anahtar kelimeler:

Özet:

 Financial time-series data are nonlinear, complex, influenced by many economic factors, and are difficult to predict. Several traditional statistical methods have been developed for financial time series modeling. However, because it is now easier to record, analyze, and transform big data into meaningful information, the use of machine learning algorithms in financial forecast development has increased in recent years. In this study, the data of EREGL stocks, which are among the stocks traded in the main metal market in the Borsa İstanbul index, are analyzed using time series methods and then modeled using ARIMA and deep models. In the developed deep learning method, the prediction performance improved with data preprocessing stages, feature extraction studies, and different time windows. For deep learning algorithms to be used in time-series studies, a framework of time delays must be used. In this study, scenarios for different time delays and performance comparisons are performed between ARIMA models and deep learning models using long-short term memory LSTM, gated repeating unit GRU, and recursive neural network RNN algorithms. Experimental studies demonstrate that the RNN algorithm has a better prediction performance than the others and predicts with an average accuracy of 93 on the test dataset. Key Words: ARIMA, BIST, Deep Learning, GRU, LSTM, RNN, Stock Price Prediction JEL Classification: E47, G17, E37.

 Financial time-series data are nonlinear, complex, influenced by many economic factors, and are difficult to predict. Several traditional statistical methods have been developed for financial time series modeling. However, because it is now easier to record, analyze, and transform big data into meaningful information, the use of machine learning algorithms in financial forecast development has increased in recent years. In this study, the data of EREGL stocks, which are among the stocks traded in the main metal market in the Borsa İstanbul index, are analyzed using time series methods and then modeled using ARIMA and deep models. In the developed deep learning method, the prediction performance improved with data preprocessing stages, feature extraction studies, and different time windows. For deep learning algorithms to be used in time-series studies, a framework of time delays must be used. In this study, scenarios for different time delays and performance comparisons are performed between ARIMA models and deep learning models using long-short term memory LSTM, gated repeating unit GRU, and recursive neural network RNN algorithms. Experimental studies demonstrate that the RNN algorithm has a better prediction performance than the others and predicts with an average accuracy of 93 on the test dataset. Key Words: ARIMA, BIST, Deep Learning, GRU, LSTM, RNN, Stock Price Prediction JEL Classification: E47, G17, E37.

 Financial time-series data are nonlinear, complex, influenced by many economic factors, and are difficult to predict. Several traditional statistical methods have been developed for financial time series modeling. However, because it is now easier to record, analyze, and transform big data into meaningful information, the use of machine learning algorithms in financial forecast development has increased in recent years. In this study, the data of EREGL stocks, which are among the stocks traded in the main metal market in the Borsa İstanbul index, are analyzed using time series methods and then modeled using ARIMA and deep models. In the developed deep learning method, the prediction performance improved with data preprocessing stages, feature extraction studies, and different time windows. For deep learning algorithms to be used in time-series studies, a framework of time delays must be used. In this study, scenarios for different time delays and performance comparisons are performed between ARIMA models and deep learning models using long-short term memory LSTM, gated repeating unit GRU, and recursive neural network RNN algorithms. Experimental studies demonstrate that the RNN algorithm has a better prediction performance than the others and predicts with an average accuracy of 93 on the test dataset. Key Words: ARIMA, BIST, Deep Learning, GRU, LSTM, RNN, Stock Price Prediction JEL Classification: E47, G17, E37.

 Finansal zaman serisi verileri doğrusal olmayan, karmaşık, birçok ekonomik faktörden etkilenen ve tahmin edilmesi zor verilerdir. Çok boyutlu ilişkilerin tahminini gerektiren finansal zaman serisi modelleri için çeşitli istatistiksel yöntemler geliştirilmiştir. Ancak günümüzde büyük verilerin kaydedilmesi, analiz edilmesi ve anlamlı bilgiye dönüştürülmesi kolaylaştığından dolayı finansal tahmin geliştirmede makine öğrenmesi algoritmalarının kullanımı özellikle son yıllarda artmıştır. Bu çalışmada, Borsa İstanbul endeksinde metal ana pazarında işlem gören EREGL hissesine ait veriler zaman serisi yöntemleri ile analiz edilmiş ardından ARIMA ve derin öğrenme modelleri ile tahmin edilmiştir. Geliştirilen derin öğrenme yönteminde veri ön işleme aşamaları, özellik çıkarımı çalışmaları ve farklı zaman çerçeveleri ile tahmin performansı iyileştirilmiştir. Derin öğrenme algoritmalarının zaman serisi çalışmalarında kullanılabilmesi için zaman gecikmelerinden oluşan bir çerçeve kullanılmalıdır. Bu çalışmada, farklı zaman gecikmeleri için senaryolar denenmiş ve performans kıyaslaması ARIMA modelleri ve uzun-kısa vadeli bellek LSTM, geçitli tekrarlayan ünite GRU ve özyineli sinir ağları RNN algoritmalarını kullanan derin öğrenme modelleri arasında gerçekleştirilmiştir. Deneysel çalıştırmalar ile RNN algoritmasının diğerlerine göre daha iyi tahmin performansına sahip olduğu ve ele alınan test veri seti üzerinde ortalama 93lük doğrulukla tahmin ettiği ortaya konulmuştur. Anahtar Kelimeler: ARIMA, BIST, Derin Öğrenme, GRU, Hisse Senedi Tahmini, LSTM, RNN JEL Sınıflandırması: E47, G17, E37.

Makale detayları

Performance Comparisons of Deep Learning and ARIMA: A Borsa Istanbul Stock Example

Yazarlar:

Caner Erden

Yayıncı:

Manisa Celal Bayar University

Ana Başlıklar:

Mali Tablo Analizi

Dergi:

Yönetim ve Ekonomi Dergisi Vol.30 Issue 3, pp.419-438

Belge Tipi:

Makale

Yayın Tarihi:

9/2023

Belge Dili:

Türkçe

Document ID (Org No.):

2001298303

pISSN:

1302-0064

BIST

LSTM

RNN

GRU

ARIMA

Stock Price Prediction

Deep LearningHisse Senedi Tahmini

BIST

LSTM

RNN

GRU

ARIMA

Derin Öğrenme

Sayı

Makale detayları

Yazarlar: Caner Erden

Yayıncı: Manisa Celal Bayar University

Ana Başlıklar: Mali Tablo Analizi

Dergi: Yönetim ve Ekonomi Dergisi Vol.30 Issue 3, pp.419-438

Belge Tipi: Makale

Yayın Tarihi: 9/2023

Belge Dili: Türkçe

Document ID (Org No.): 2001298303

pISSN: 1302-0064

Sayı

Yazarlar:

Caner Erden

Yayıncı:

Manisa Celal Bayar University

Ana Başlıklar:

Mali Tablo Analizi

Dergi:

Yönetim ve Ekonomi Dergisi Vol.30 Issue 3, pp.419-438

Belge Tipi:

Makale

Yayın Tarihi:

9/2023

Belge Dili:

Türkçe

Document ID (Org No.):

2001298303

pISSN:

1302-0064