BAYES AĞ MODELLERİ İLE HİSSE SENEDİ GETİRİLERİNİN KARŞILIKLI DİNAMİK İLİŞKİLERİ

Yıl 2018, , 709 - 720, 20.01.2018

Fatma Busem Hatipoğlu Umut Uyar

https://doi.org/10.18092/ulikidince.346501

Öz

Modern
portföy teorisinde, portföyde yer alan menkul kıymetler arasındaki ilişkinin
yönünün ve derecesinin riskin azaltılması yönünde etkili olduğu
belirtilmektedir (Markowitz, 1952). Teoride, birbirleriyle yüksek korelasyon
içinde bulunan menkul kıymetlerin aynı portföyde yer almasından
kaçınılmaktadır. Ancak korelasyon katsayısı, iki rassal değişken arasındaki
doğrusal ilişkinin yönünü ve gücünü belirtmektedir. Bayes ağlar kullanılarak
oluşturulan modeller menkul kıymetler arasındaki olasılıksal ilişkiyi görsel
olarak sunabilmekte ve yeni bilgi geldiğinde ağda yer alan menkul kıymet getiri
değerleri eşzamanlı olarak güncellenebilmektedir. Çalışmanın amacı, 2011-2016
dönemleri arasında Borsa İstanbul Ulusal-100 (BIST-100) endeksinde kesintisiz
faaliyet gösteren hisse senetlerine ait getirilerin birbirleri ile olan ilişkilerini
bir makine öğrenmesi olan Bayes ağ modelleri kullanarak araştırmaktır.
Çalışmada Bayes ağ modelleri kullanılarak elde edilecek detaylı ilişkiler ile
yatırımcıların portföy seçimlerinde kullanabilecekleri nitel ve nicel bilgiler
yer almaktadır.

Anahtar Kelimeler

Makine Öğrenmesi, Bayes Ağlar, Portföy Seçim Teorisi

THE EXAMINE OF DYNAMIC INTERRELATION OF STOCK RETURNS: BAYESIAN NETWORK MODELS

Öz

In modern portfolio theory, it is stated that the
relationship between the securities in the portfolio is influenced by the
direction and degree of risk reduction (Markowitz, 1952). In theory, securities
that are highly correlated with each other are avoided from being placed in the
same portfolio. However, the correlation coefficient indicates the direction
and power of the linear relationship between the two random variables. Models
created using Bayesian networks can visually present the probabilistic
relationship between securities, and when new information is available, the
securities return values ​​in the network can be updated simultaneously. The
aim of the study is to investigate the relationships between stocks that have
been operating continuously in the Stock Exchange Istanbul National-100
(BIST-100) index between 2011-2016 by using Bayes network models which are
machine learning. In the study, detailed relationships to be obtained by using
Bayesian network models and qualitative and quantitative information that
investors can use in portfolio selection are included.

Anahtar Kelimeler

Machine Learning, Bayesian Networks, Portfolio Selection Theory

Kaynakça

• Bensi, M. T., Kiureghian, A. D. ve Straub, D. (2011). A Bayesian Network Methodology for Infrastructure Seismic Risk Assessment and Decision Support (Rapor No. PEER Report 2011/02). Pacific Earthquake Engineering Research Center. Erişim adresi http://peer.berkeley.edu/publications/peer_reports/reports_2011/webPEER-2011-02-BENSIetal.pdf
• Conrady, S. ve Jouffe, L. (2015). Bayesian Networks & BayesiaLab–A Practical Introduction for Researchers, Bayesia USA.
• Demirer, R., Mau, R. ve Shenoy, C. (2006). Bayesian Networks: A Decision Tool to Improve Portfolio Risk Analysis. Journal of Applied Finance, (16: 2), 106-133.
• Greppi, A. (2014, June-July). A Bayesian Network Approach to Portfolio Management, First DREAMT Research in Progress Workshop, Pavia. Erişim adresi https://www.researchgate.net/profile/Fabio_Stella/publication/299760529_Bayesian_Networks_for_Portfolio_Analysis_and_Optimization/links/574ff94908ae1880a8229222/Bayesian-Networks-for-Portfolio-Analysis-and-Optimization.pdf
• Greppi, A., De Giuli M. E. ve Tarantola C. (2013). Bayesian Network for Stock Picking. Erişim Adresi http://convegni.unica.it/cladag2015/files/2015/10/Greppi.pdf
• Hoe, T. K. (2014). A Machine Learning-based Decision Support Tools for Portfolio Risk Analysis. (Yayımlanmamış Yüksek Lisans Tezi). Tunku Abdul Rahman Üniversitesi, Malaysia.
• Lauria, E. J. M. ve Duchessi, P. (2007). A methodology for developing Bayesian networks: An application to information technology (IT) implementation, European Journal of Operational Research, 179, 234-252.
• Markowitz, H. (1952). Portfolio Selection, The Journal of Finance, Blackwell Publishing, Vol: 7, No: 1, 77-91.
• Mittal, A., Kassim, A. ve Tan, T. (2007). Bayesian Network Technologies: Applications and Graphical Models, IGI Global, USA.
• Nagarajan, R., Scutari, M. ve Lebre, S. (2013). Bayesian Networks in R with Applications in Systems Biology, Springer Science+Business Media, New York.
• Olbryś, J. (2009). Forecasting Portfolio Return Based on Bayesian Network Model, [in:] W. Milo, G. Szafrański, P. Wdowiński (eds.) Financial Markets. Principles of Modelling, Forecasting, and Decision-Making, FindEcon Monograph Series: Advances in Financial Market Analysis, Vol. 7, Lodz University Press, 157-171.
• Sammut, C. ve Webb, G. I. (2011). Encyclopedia of Machine Learning, Springer.
• Shenoy, C. ve Shenoy, P. P. (1998). Bayesian Networks: A Decision Tool to Improve Portfolio Risk Analysis (Working paper). School of Business, University of Kansas. Erişim adresi http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.111.1084&rep=rep1&type=pdf
• Sucar, L. E. (2015). Probabilistic Graphical Models Principles and Applications, Springer-Verlag, London.
• Trucco, P., Cagno, E., Ruggeri, F. ve Grande, O. (2008). Bayesian Belief Network modeling of organizational factors in risk analysis: A case study in maritime transportation, Reliability Engineering and System Safety, 93, 823-834.
• Tseng, C. (2003). Comparing Artificial Intelligence Systems for Stock Portfolio Selection. Erişim adresi depts.washington.edu/sce2003/Papers/236.pdf. (24.07.2017)
• Villa, S. ve Stella, F. (2012). Bayesian Networks for Portfolio Analysis and Optimization, Financial Decision Making Using Computational Intelligence, Vol. 70, 209-232.
• Zuo, Y. ve Kita, E. (2012). Up/Down Analysis of Stock Index by Using Bayesian Network, Engineering Management Research, Vol. 1, No. 2. 46-52.