Araştırma Makalesi


DOI :10.26650/ekoist.2021.35.180033   IUP :10.26650/ekoist.2021.35.180033    Tam Metin (PDF)

Türkiye’nin Konut Satışı Değerlerinin Yapay Sinir Ağları İle Öngörülmesi

Burcu Yaman Selçi

Gayrimenkul sektöründe konut arz ve talebinin dengede tutulabilmesi için konut satış tahminlerinin güçlü tahminler yapabilecek bir analiz yöntemi ile doğru bir şekilde yapılması büyük önem taşır. Fakat literatürde konut satış tahminlerine odaklanan çalışma sayısının oldukça az olduğu ve yeni nesil tekniklerden yapay sinir ağları ile tahmin yapan çalışma sayısının kısıtlılığı dikkat çekicidir. Bu nedenle bu çalışmanın amacı Türkiye’de konut satışlarının tahminini ve öngörüsünü yapay sinir ağları ile gerçekleştirerek literatüre katkı sunmaktır. Çalışmada konut-fiyat endeksi, yeni konut-fiyat endeksi, yeni olmayan konut-fiyat endeksi, yabancılara yapılan konut satışı, bankalarca TL üzerinden konut kredilerine açılan faiz oranları, tüketici fiyat endeksi ve kur bağımsız değişkenler olarak seçilmiş ve konut satışı bağımlı değişken olarak kullanılarak yapay sinir ağlarında bir model geliştirilmiştir. Veriler 2013:01-2019:12 dönemlerini kapsayacak şekilde aylık olarak alınmış ve analizler MATLAB R2013a programında gerçekleştirilmiştir. Tahmin ve öngörü analizi için NARX ağı kullanılarak 2013:01-2019:12 döneminin tahmini ve 2020:01 döneminin öngörüsü elde edilmiştir. Performans ölçütü olarak ise MSE kullanılmıştır. Analiz sonucunda yapay sinir ağlarının ürettiği tahmin değerlerinin ve 2020:01 dönemine ait öngörü değerinin gerçek değerler ile oldukça yakın olduğu ve yapay sinir ağlarının mevsimlik etkileri saptayabildiği tespit edilmiştir. MSE değerinin küçüklüğü de tahmin ve öngörü başarısını ortaya koymuştur. Bu durum yapay sinir ağlarının konut satışı tahmininde ve öngörüsünde güçlü istatistiksel sonuçlar ürettiğini doğrular niteliktedir.

JEL Classification : C45 , C51 , C53
DOI :10.26650/ekoist.2021.35.180033   IUP :10.26650/ekoist.2021.35.180033    Tam Metin (PDF)

Prediction Using Artificial Neural Network of Turkey's Housing Sales Value

Burcu Yaman Selçi

In order to keep the supply and demand in the real estate sector in balance, it is very important to make accurate estimates of house sales with an analysis method that can make strong predictions. However, it is noteworthy that the number of studies focusing on house sales estimates in the literature is quite low and the number of studies that make predictions with artificial neural networks from new generation techniques is remarkable. Therefore the aim of this study is to contribute to the prediction and forecasting of sales literature houses in Turkey performing with artificial neural networks. In the study, housing-price index, new housing-price index, non-new housing-price index, house sales to foreigners, interest rates opened to housing loans over TL, consumer price index and exchange rate were selected as independent variables and residential sales were used as dependent variables. A model has been developed in neural networks. The data were taken monthly to cover the periods of 2013: 01-2019: 12 and the analyzes were carried out in the MATLAB R2013a program. Using the NARX network for prediction and forecasting analysis, the prediction of 2013: 01- 2019: 12 period and the prediction of 2020: 01 period was obtained. MSE was used as a performance criterion. As a result of the analysis, it has been determined that the predicted values produced by artificial neural networks and the predictive value of 2020: 01 are quite close to real values and artificial neural networks can detect seasonal effects. The smallness of the MSE value also proved the success of forecasting and forecasting. This confirms that artificial neural networks produce strong statistical results in predicting and predicting residential sales.

JEL Classification : C45 , C51 , C53

GENİŞLETİLMİŞ ÖZET


Especially in recent years, it is seen that the real estate sector has an important place in the economy and affects macroeconomic balances. Housing supply and demand are affected by seasonal factors and affect the market. For this reason, it is very important to make the estimations for the real estate sector by choosing methods that produce strong statistical results that can capture the seasonal effects. Artificial neural networks are at the top of these powerful statistical techniques. Artificial neural networks are frequently preferred by researchers in the prediction of economic data and prospective forecasts especially in recent years. It can be said that the reason why artificial neural networks are preferred in contrast to traditional systems is the advantages offered to researchers. At the beginning of these advantages, it can be said that artificial neural networks can work with an unlimited number of variables and parameters, and can make generalizations and predictions with the help of the data shown to it. Thanks to these advantages, it can be said that artificial neural networks can be frequently used in the estimation of economic data in the coming years.

In this study, artificial neural networks have been preferable due to superior features and Turkey’s estimated sales value and forecasts of housing was conducted. Monthly data covering the periods of 2013: 01-2019: 12 were analyzed in the MATLAB R2013a program. As a result of the research conducted in the literature for the analysis, housing-price index, new housing-price index, non-new housingprice index, housing sales to foreigners, interest rates opened by banks to housing loans over TL, consumer price index and exchange rate were selected as independent variables and sales of housing. used as a dependent variable. NARX (Nonlinear Autoregressive Exogeneus) network was first trained with the Levenberg-Marquardt back-propagation algorithm (trainlm) in the model, which has ten independent variables, one dependent variable and 924 data in total. 90% of the data was randomly allocated for education, 5% for validity and 5% for testing, as it gave the best results as a result of the trials. The number of hidden layers is chosen as 10, while the number of lags is taken as 5. Mean Squared Error (MSE) was chosen as the performance criterion.

As a result of the analysis, since the NARX network uses historical values in the studies to be predicted and the number of delays is taken as 5, the estimation of the first five values in the data set has been estimated. When the results produced by the neural network in Turkey real value of residential sales values compared to each other, which is very close and has been found to be quite small measure of performance of the value of MSE. In the second stage of the analysis prediction of Turkey’s housing sales value for the period of 2020:01 were made. When the real value of the period related to the value produced by artificial neural networks as a result of the prediction is compared, it has been determined that the predictive performance of artificial neural networks is also quite good. The results of the analysis confirm that the artificial neural networks can achieve seasonal effects in predicting and forecasting the house sales and producing strong statistical results in accordance with the literature. Although the results of the study are expected to contribute to the literature, the study has some limitations. This gives the number of studies in order to overcome the limitations, frequency and number of arguments can be increased and a traditional time series method can also be included in the study a model comparison made Turkey’s housing sales Values, estimates that the most powerful model can be determined. 


PDF Görünüm

Referanslar

  • Bakhary, N., Yahya, K., & Ng, C. N. (2004). Univariate Artificial Neural Network in Forecasting Demand of Low Cost House in Petaling Jaya. Jurnal Teknologi, 40(1), 1-16. google scholar
  • Boussaada, Z., Curea, O., Remaci, A., Camblong, H., & Mrabet Bellaaj, N. (2018). A nonlinear autoregressive exogenous (NARX) neural network model for the prediction of the daily direct solar radiation. Energies, 11(3), 620. google scholar
  • Chaudhuri, T. D., & Ghosh, I. (2016). Artificial Neural Network and Time Series Modeling Based Approach To Forecasting The Exchange Rate İn A Multivariate Framework. arXiv preprint arXiv:1607.02093. google scholar
  • Dahan, A. A. (2018). The Future of The Real Estate Industry of Dubai: The Demand for Real Estates. Journal of Global Economic. google scholar
  • GOH, B. H. (1998). Forecasting Residential Construction Demand in Singapore: A Comparative Study of The Accuracy of Time Series, Regression and Artificial Neural Network Techniques. Engineering, Construction and Architectural Management. google scholar
  • Haykin, S. (2009). Neural Networks and Learning Machines. (3rd Ed.). New York: Prentice Hall. google scholar
  • Karaaslan, A., & Özden, K. Ö. (2016). Housing Demand in Turkey: Application of Grey Forecasting Model. Mediterranean Journal of Social Sciences, 7(2), 52. google scholar
  • Khalafallah, A. (2008). Neural Network Based Model for Predicting Housing Market Performance. Tsinghua Science and Technology, 13(S1), 325-328. google scholar
  • Lin, T., Horne, B. G., Tino, P., & Giles, C. L. (1996). Learning Long-Term Dependencies in NARX Recurrent Neural Networks. IEEE Transactions on Neural Networks, 7(6), 1329-1338. google scholar
  • Mathworks, 2018. Design Time Series NARX Feedback Neural Networks. https://www.mathworks.com/help/deeplearning/ug/design-time-series-narx-feedback-neural-networks.html#responsive_offcanvas (Erişim Tarihi: 21.08.2020). google scholar
  • Öztemel, E. (2006). Yapay Sinir Ağları. (2. Baskı). İstanbul: Papatya Yayıncılık. google scholar
  • Temür, A. S., Akgün, M., & Temür, G. (2019). Predicting Housing Sales in Turkey Using ARIMA, LSTM and Hybrid Models. Journal of Business Economics and Management, 20(5), 920-938. google scholar
  • Yavuz, E. (2018). Yapay Sinir Ağı Kullanarak Kontrol Alan Ağları İçin Çevrim İçi Mesaj Zamanlaması Optimizasyonu. Süleyman Demirel Üniversitesi, Isparta. google scholar
  • Yıldız, B. (2009). Finansal Analizde Yapay Zeka. Ankara: Detay Yayıncılık. google scholar
  • Yılmaz, B. (2015). Akarçay Havzasında Çözünmüş Oksijen Değerlerinin Yapay Sinir Ağları İle Belirlenmesi. Uzmanlık Tezi, Ankara. google scholar
  • Yurtoğlu, H. (2005). Yapay Sinir Ağları Metodolojisi İle Öngörü Modellemesi: Bazı Makroeknomik Değişkenler İçin Türkiye örneği. DPT. google scholar
  • Zainun, N. Y. (2011). Computerized Model to Forecast Low-Cost Housing Demand in Urban Area in Malaysia Using Artificial Neural Networks (ANN) (Doctoral dissertation, Loughborough University). google scholar
  • https://evds2.tcmb.gov.tr/ (Erişim Tarihi: 01.03.2020). google scholar

Atıflar

Biçimlendirilmiş bir atıfı kopyalayıp yapıştırın veya seçtiğiniz biçimde dışa aktarmak için seçeneklerden birini kullanın


DIŞA AKTAR



APA

Yaman Selçi, B. (2021). Türkiye’nin Konut Satışı Değerlerinin Yapay Sinir Ağları İle Öngörülmesi. Ekoist: Journal of Econometrics and Statistics, 0(35), 19-32. https://doi.org/10.26650/ekoist.2021.35.180033


AMA

Yaman Selçi B. Türkiye’nin Konut Satışı Değerlerinin Yapay Sinir Ağları İle Öngörülmesi. Ekoist: Journal of Econometrics and Statistics. 2021;0(35):19-32. https://doi.org/10.26650/ekoist.2021.35.180033


ABNT

Yaman Selçi, B. Türkiye’nin Konut Satışı Değerlerinin Yapay Sinir Ağları İle Öngörülmesi. Ekoist: Journal of Econometrics and Statistics, [Publisher Location], v. 0, n. 35, p. 19-32, 2021.


Chicago: Author-Date Style

Yaman Selçi, Burcu,. 2021. “Türkiye’nin Konut Satışı Değerlerinin Yapay Sinir Ağları İle Öngörülmesi.” Ekoist: Journal of Econometrics and Statistics 0, no. 35: 19-32. https://doi.org/10.26650/ekoist.2021.35.180033


Chicago: Humanities Style

Yaman Selçi, Burcu,. Türkiye’nin Konut Satışı Değerlerinin Yapay Sinir Ağları İle Öngörülmesi.” Ekoist: Journal of Econometrics and Statistics 0, no. 35 (Jul. 2022): 19-32. https://doi.org/10.26650/ekoist.2021.35.180033


Harvard: Australian Style

Yaman Selçi, B 2021, 'Türkiye’nin Konut Satışı Değerlerinin Yapay Sinir Ağları İle Öngörülmesi', Ekoist: Journal of Econometrics and Statistics, vol. 0, no. 35, pp. 19-32, viewed 6 Jul. 2022, https://doi.org/10.26650/ekoist.2021.35.180033


Harvard: Author-Date Style

Yaman Selçi, B. (2021) ‘Türkiye’nin Konut Satışı Değerlerinin Yapay Sinir Ağları İle Öngörülmesi’, Ekoist: Journal of Econometrics and Statistics, 0(35), pp. 19-32. https://doi.org/10.26650/ekoist.2021.35.180033 (6 Jul. 2022).


MLA

Yaman Selçi, Burcu,. Türkiye’nin Konut Satışı Değerlerinin Yapay Sinir Ağları İle Öngörülmesi.” Ekoist: Journal of Econometrics and Statistics, vol. 0, no. 35, 2021, pp. 19-32. [Database Container], https://doi.org/10.26650/ekoist.2021.35.180033


Vancouver

Yaman Selçi B. Türkiye’nin Konut Satışı Değerlerinin Yapay Sinir Ağları İle Öngörülmesi. Ekoist: Journal of Econometrics and Statistics [Internet]. 6 Jul. 2022 [cited 6 Jul. 2022];0(35):19-32. Available from: https://doi.org/10.26650/ekoist.2021.35.180033 doi: 10.26650/ekoist.2021.35.180033


ISNAD

Yaman Selçi, Burcu. Türkiye’nin Konut Satışı Değerlerinin Yapay Sinir Ağları İle Öngörülmesi”. Ekoist: Journal of Econometrics and Statistics 0/35 (Jul. 2022): 19-32. https://doi.org/10.26650/ekoist.2021.35.180033



ZAMAN ÇİZELGESİ


Gönderim30.05.2020
Kabul04.11.2020
Çevrimiçi Yayınlanma31.12.2021

LİSANS


Attribution-NonCommercial (CC BY-NC)

This license lets others remix, tweak, and build upon your work non-commercially, and although their new works must also acknowledge you and be non-commercial, they don’t have to license their derivative works on the same terms.


PAYLAŞ




İstanbul Üniversitesi Yayınları, uluslararası yayıncılık standartları ve etiğine uygun olarak, yüksek kalitede bilimsel dergi ve kitapların yayınlanmasıyla giderek artan bilimsel bilginin yayılmasına katkıda bulunmayı amaçlamaktadır. İstanbul Üniversitesi Yayınları açık erişimli, ticari olmayan, bilimsel yayıncılığı takip etmektedir.