Hakan GÜNDÜZ, Yusuf YASLAN, Zehra CATALTEPE

Stock daily return prediction using expanded features and feature selection

Stock market prediction is a very noisy problem and the use of any additional information to increase accuracy is necessary. In this paper, for the stock daily return prediction problem, the set of features is expanded to include indicators not only for the stock to be predicted itself but also a set of other stocks and currencies. Afterwards, different feature selection and classi cation methods are utilized for prediction. The daily close returns of the 3 most traded stocks (GARAN, THYAO, and ISCTR) in Borsa _ Istanbul (BIST) are predicted using indicators computed on those stocks, indicators for all the other stocks listed in the BIST100 index, and indicators on the dollar-gold prices. Twenty- ve different indicators on daily stock prices are computed to form feature vectors for each trading day. These feature vectors are assigned class labels according to the daily close returns. Expanding the feature space with BIST100 stocks features results in a high dimensional feature space, with possibly noisy or irrelevant features. Therefore, feature selection methods are utilized to select the most informative features. In order to determine relevance scores of features, fast lter-based methods, gain ratio and relief, are used. Experiments are performed based on individual stock features, dollar-gold features (DG), BIST100 stock features (BIST100), and a combination of BIST100 and DG with and without feature selection. Using the gain ratio feature selection with a gradient boosting machine (GBM), the movements of GARAN stock were predicted with an accuracy of 0.599 and an F-measure of 0.614. For THYAO, the relief feature selection with the GBM gave an accuracy of 0.558, and for ISCTR, the gain ratio feature selection with logistic regression achieved an accuracy of 0.581. It was found that using BIST100 stock features boosts classi cation results for all stocks in terms of accuracy.

PDF

___

[1] Abu-Mostafa YS, Atiya AF. Introduction to nancial forecasting. Appl Intell 1996; 6: 205-213.
[2] Marcek D. Stock price forecasting: statistical, classical and fuzzy neural network approach. In: MDAI; 2004. pp. 41-48.
[3] Atsalakis GS, Valavanis KP. Surveying stock market forecasting techniques{Part II: Soft computing methods. Expert Syst Appl 2009; 36: 5932-5941.
[4] Gunduz H, Cataltepe Z. Borsa Istanbul (BIST) daily prediction using nancial news and balanced feature selec- tion. Expert Syst Appl 2015; 42: 9001-9011.
[5] Guresen E, Kayakutlu G, Daim TU. Using arti cial neural network models in stock market index prediction. Expert Syst Appl 2011; 38: 10389-10397.
[6] Alkhatib K, Najadat H, Hmeidi I, Shatnawi MKA. Stock price prediction using k-nearest neighbor (kNN) algo- rithm. International Journal of Business, Humanities and Technology 2013; 3: 32-44.
[7] Ballings M, Van den Poel D, Hespeels N, Gryp R. Evaluating multiple classi ers for stock price direction predic- tion. Expert Syst Appl 2015; 42: 7046-7056.
[8] Patel J, Shah S, Thakkar P, Kotecha K. Predicting stock and stock price index movement using trend deterministic data preparation and machine learning techniques. Expert Syst Appl 2015; 42: 259-268.
[9] Pehlivanli AC, Asikgil B, Gulay G. Indicator selection with committee decision of lter methods for stock market price trend in ISE. Appl Soft Comput 2016; 49: 792-800.
[10] Khaidem L, Snehanshu S, Sudeepa RD. Predicting the direction of stock market prices using random forest. arXiv 2016; 1605.00003.
[11] Ergur B, Cataltepe Z. Relational stock market network analysis. In: 21st Signal Processing and Communications Applications Conference (SIU); 2013.
[12] Basti E, Kuzey C, Delen D. Analyzing initial public offerings' short-term performance using decision trees and SVMs. Decis Support Syst 2015; 73: 15-27.
[13] Dutta A, Bandopadhyay G, Sengupta S. Prediction of stock performance in Indian stock market using logistic regression. International Journal of Business and Information 2015; 7: 105-136.
[14] Martinez LC, da Hora DN, Palotti JRDM, Meira W, Pappa GL. From an arti cial neural network to a stock market day-trading system: a case study on the BM&F BOVESPA. In: IJCNN; 2009. pp. 2006-2013.
[15] Kayal A. A neural networks ltering mechanism for foreign exchange trading signals. In: Intelligent Computing and Intelligent Systems; 2010.
[16] Kim K. Financial time series forecasting using support vector machines. Neurocomputing 2003; 55: 307-319.
[17] Markovic IP, Stojanovic MB, Stankovic JZ, Bozic MM. Stock market trend prediction using support vector ma- chines. Facta Universitatis Series Automatic Control and Robotics 2014; 13: 147-158.
[18] Weng B, Ahmed MA, Megahed FM. Stock market one-day ahead movement prediction using disparate data sources. Expert Syst Appl 2017; 79: 153-163.
[19] Barak S, Modarres M. Developing an approach to evaluate stocks by forecasting effective features with data mining methods. Expert Syst Appl 2015; 42: 1325-1339.
[20] Zhong X, Enke D. Forecasting daily stock market return using dimensionality reduction. Expert Syst Appl 2017; 67: 126-139.
[21] Cavalcante RC, Oliveira ALI. An autonomous trader agent for the stock market based on online sequential extreme learning machine ensemble. In: IJCNN; 2014.
[22] Barak S, Arjmand A, Ortobelli S. Fusion of multiple diverse predictors in stock market. Inform Sciences 2017; 36: 90-102.
[23] Dey S, Kumar Y, Saha S, Basak S. Forecasting to classi cation: predicting the direction of stock market price using Xtreme Gradient Boosting. Working paper. DOI: 10.13140/RG.2.2.15294.48968.
[24] Sosvilla-Rivero S, Rodrguez PN. Linkages in international stock markets: evidence from a classi cation procedure. Appl Econ 2010; 42: 2081-2089.
[25] Bildirici M, Ersin O O. Improving forecasts of GARCH family models with the arti cial neural networks: an application to the daily returns in Istanbul Stock Exchange. Expert Syst Appl 2009; 36: 7355-7362.
[26] Kara Y, Boyacioglu MA, Baykan OK. Predicting direction of stock price index movement using arti cial neural networks and support vector machines: The sample of the Istanbul Stock Exchange. Expert Syst Appl 2011; 38: 5311-5319.
[27] Kumar D, Meghwani SS, Thakur M. Proximal support vector machine based hybrid prediction models for trend forecasting in nancial markets. Journal of Computational Science 2016; 17: 1-13.
[28] Quinlan JR. Improved use of continuous attributes in C4.5. J Artif Intell Res 1996; 4: 77-90.
[29] Kira K, Rendell LA. The feature selection problem: traditional methods and a new algorithm. In: AAAI; 1992. pp. 129-134.
[30] Hosmer Jr DW, Lemeshow S, Sturdivant RX. Applied Logistic Regression (Vol. 398). New York, NY, USA: John Wiley & Sons, 2013.
[31] Freund Y, Schapire RE. Experiments with a new boosting algorithm. In: ICML (Vol. 96); 1996. pp. 148-156.
[32] Schapire RE, Freund Y. Boosting: Foundations and Algorithms. Cambridge, MA, USA: MIT Press, 2012.
[33] Friedman JH. Greedy function approximation: a gradient boosting machine. Ann Stat 2001; 29: 1189-1232.
[34] Chen T, Guestrin C. Xgboost: A scalable tree boosting system. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining; 2016. pp. 785-794.
[35] Jadhav S, He H, Jenkins KW. An academic review: applications of data mining techniques in nance industry. In- ternational Journal of Soft Computing and Arti cial Intelligence 2017; 4: 79-95.
[36] Fersini E, Messina E, Pozzi FA. Sentiment analysis: Bayesian ensemble learning. Decis Support Syst 2014; 68: 26-38.
[37] Nagy GI, Barta G, Kazi S, Borbely G, Simon G. GEFCom2014: Probabilistic solar and wind power forecasting using a generalized additive tree ensemble approach. Int J Forecasting 2016; 32: 1087-1093.