Abstract
In this study, we generate 50 Multi-layer Perceptons, 50 Radial Basis Functions, 50 Higher Order Neural Networks and 50 Recurrent Neural Network and we explore their utility in forecasting and trading the DJIA, NASDAQ 100 and the NIKKEI 225 stock indices. The statistical significance of the forecasts is examined through the False Discovery Ratio of Bajgrowicz and Scaillet (J Financ Econ 106(3):473–491, 2012). Two financial everages, based on the levels of financial stress and the financial volatility respectively, are also applied. In terms of the results, we note that RNN have the higher percentage of significant models and present the stronger profitability compared to their Neural Network counterparts. The financial leverages doubles the trading performance of our models.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Backpropagation is the most common learning algorithm in MLPs and the most commonly used type in financial time series forecasting (Kaastra and Boyd 1996).
As risk free rate we consider the effective federal funds rate. This is the interest rate at which depository institutions trade federal funds (balances held at Federal Reserve Banks) with each other overnight.
Stock indices can be traded with three separate ways. A trader can buy or short all stocks that constitute the index, however this might lead to substantial transaction costs and liquidity issues. The second way is through futures and the third is through an Exchange Traded Fund (ETF) on the index. The transaction costs considered in this study are an approximation of the costs that large institutional investors will be facing if they decided to trade the stock indices under study with futures or ETFs.
The finance literature is rich with financial stress indices. In our application, we need an index that is updated on a daily basis and covers USA and Japan. The OFR has these two characteristics. However, it should be noted that in the other advanced economies are included not only Japan but also the Eurozone countries. Although the fit is not perfect for this case, it is the closest to the best of our knowledge. Other stress indices have lower frequency or focus on a single economy.
A brief description of the RiskMetrics volatility model is provided in "Appendix".
References
Adeodato, P., Arnaud, A., Vasconcelos, G., Cunha, R., & Monteiro, D. (2011). MLP ensembles improve long term prediction accuracy over single networks. International Journal of Forecasting, 27(3), 661–671.
Andreou, P. C., Charalambous, C., & Martzoukos, S. H. (2008). Pricing and trading European options by combining artificial neural networks and parametric models with implied parameters. European Journal of Operational Research, 185(3), 1415–1433.
Bagheri, A., Peyhani, H. M., & Akbari, M. (2014). Financial forecasting using ANFIS networks with quantum-behaved particle swarm optimization. Expert Systems with Applications, 41(14), 6235–6250.
Bahrammirzaee, A. (2010). A comparative survey of artificial intelligence applications in finance: Artificial neural networks, expert system and hybrid intelligent systems. Neural Computing and Applications, 19(8), 1165–1195.
Bajgrowicz, P., & Scaillet, O. (2012). Technical trading revisited: False discoveries, persistence tests, and transaction costs. Journal of Financial Economics, 106(3), 473–491.
Bekiros, S. D., & Georgoutsos, D. A. (2008). Direction-of-change forecasting using a volatility-based recurrent neural network. Journal of Forecasting, 27(5), 407–417.
Bertolini, L. (2010). Trading foreign exchange carry portfolios. PhD thesis, Cass Business School, City University London.
Boston Consulting Group. (2016). The factory of the future: The ghost in the machine, Boston Consulting Group. Online report.
Broomhead, D. S., & Lowe, D. (1988). Radial basis functions, multi-variable functional interpolation and adaptive networks (No. RSRE-MEMO-4148). Royal Signals and Radar Establishment Malvern.
Cavalcante, R. C., Brasileiro, R. C., Souza, V. L., Nobrega, J. P., & Oliveira, A. L. (2016). Computational intelligence and financial markets: A survey and future directions. Expert Systems with Applications, 55, 194–211.
Chui, M., Manyika, J., & Miremadi, M. (2016). Where machines could replace humans—and where they can’t (yet). Mckinsey Quarterly, 7.
Donaldson, R. G., & Kamstra, M. (1996). Forecast combining with neural networks. Journal of Forecasting, 15(1), 49–61.
Dunis, C., Laws, J., & Sermpinis, G. (2010). Modelling and trading the EUR/USD exchange rate at the ECB fixing. The European Journal of Finance, 16(6), 541–560.
Dunis, C., & Miao, J. (2006). Volatility filters for asset management: An application to managed futures. Journal of Asset Management, 7(3–4), 179–189.
Dunis, C. L., Laws, J., & Sermpinis, G. (2011). Higher order and recurrent neural architectures for trading the EUR/USD exchange rate. Quantitative Finance, 11(4), 615–629.
Elman, J. L. (1990). Finding structure in time. Cognitive Science, 14, 179–211.
Fulcher, J., Zhang, M., & Xu, S. (2006). The application of higher-order neural networks to financial time series. In J. Kamruzzaman, R. Begg, & R. Sarker (Eds.), Artificial neural networks in finance and manufacturing. Hershey, PA: Idea Group, London.
Giles, C. L., & Maxwell, T. (1987). Learning, invariance, and generalization in high-order neural networks. Applied Optics, 26(23), 4972–4978.
Guresen, E., Kayakutlu, G., & Daim, T. U. (2011). Using artificial neural network models in stock market index prediction. Expert Systems with Applications, 38(8), 10389–10397.
Hornik, K., Stinchcombe, M., & White, H. (1989). Multilayer feedforward networks are universal approximators. Neural Networks, 2(5), 359–366.
Kaastra, I., & Boyd, M. (1996). Designing a neural network for forecasting financial and economic time series. Neurocomputing, 10(3), 215–236.
Krauss, C., Do, X. A., & Huck, N. (2017). Deep neural networks, gradient-boosted trees, random forests: Statistical arbitrage on the S&P 500. European Journal of Operational Research, 259(2), 689–702.
Lisboa, P. J. G., & Vellido, A. (2000). Business applications of neural networks, vii–xxii. In P. J. G. Lisboa, B. Edisbury, & A. Vellido (Eds.), Business applications of neural networks: The state-of-the-art of real-world applications. Singapore: World Scientific.
Lo, A. W. (2004). The adaptive markets hypothesis. Journal of Portfolio Management, 30(5), 15–29.
Maren, A. J., Harston, C. T., & Pap, R. M. (2014). Handbook of neural computing applications. Cambridge: Academic Press.
Matías, J. M., & Reboredo, J. C. (2012). Forecasting performance of nonlinear models for intraday stock returns. Journal of Forecasting, 31(2), 172–188.
Orr, M. J. L. (1996). Introduction to radial basis function networks. Centre for Cognitive Science. Scothland: Edinburgh University. http://www.anc.ed.ac.uk/&mjo/rbf.html.
Panda, C., & Narasimhan, V. (2007). Forecasting exchange rate better with artificial neural network. Journal of Policy Modeling, 29(2), 227–236.
Park, J., & Sandberg, I. W. (1991). Universal approximation using radial-basis-function networks. Neural Computation, 3(2), 246–257.
Politis, D. N., & Romano, J. P. (1994). The stationary bootstrap. Journal of the American Statistical Association, 89(428), 1303–1313.
Romano, J. P., & Wolf, M. (2007). Control of generalized error rates in multiple testing. The Annals of Statistics, 35(4), 1378–1408.
Sermpinis, G., Stasinakis, C., & Dunis, C. (2014). Stochastic and genetic neural network combinations in trading and hybrid time-varying leverage effects. Journal of International Financial Markets, Institutions and Money, 30, 21–54.
Sermpinis, G., Theofilatos, K., Karathanasopoulos, A., Georgopoulos, E. F., & Dunis, C. (2013). Forecasting foreign exchange rates with adaptive neural networks using radial-basis functions and particle swarm optimization. European Journal of Operational Research, 225(3), 528–540.
Shapiro, A. F. (2000). A Hitchhiker’s guide to the techniques of adaptive nonlinear models. Insurance, Mathematics and Economics, 26(2), 119–132.
Storey, J. D. (2002). A direct approach to false discovery rates. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 64(3), 479–498.
Tenti, P. (1996). Forecasting foreign exchange rates using recurrent neural networks. Applied Artificial Intelligence, 10(6), 567–581.
White, H. (2000). A reality check for data snooping. Econometrica, 68(5), 1097–1126.
Yang, J., Cabrera, J., & Wang, T. (2010). Nonlinearity, data-snooping, and stock index ETF return predictability. European Journal of Operational Research, 200(2), 498–507.
Yang, J., Su, X., & Kolari, J. W. (2008). Do Euro exchange rates follow a martingale? Some out-of-sample evidence. Journal of Banking & Finance, 32(5), 729–740.
Yingwei, L., Sundararajan, N., & Saratchandran, P. (1998). Performance evaluation of a sequential minimal radial basis function (RBF) neural network learning algorithm. IEEE Transactions on Neural Networks, 9(2), 308–318.
Yu, H., Xie, T., Paszczynski, S., & Wilamowski, B. M. (2011). Advantages of radial basis function networks for dynamic system design. IEEE Transactions on Industrial Electronics, 58(12), 5438–5450.
Zhang, G., Patuwo, B. E., & Hu, M. Y. (1998). Forecasting with artificial neural networks: The state of the art. International Journal of Forecasting, 14(1), 35–62.
Zhang, M. (2009). Artificial higher order neural networks for economics and business. Hershey: IGI Global.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix: RiskMetrics volatility
Appendix: RiskMetrics volatility
The RiskMetrics volatility model is one of the simpler methods to forecast the financial volatility. Nevertheless, it is popular among quantitative finance professionals. Derived from the from the GARCH(1,1) model but with fixed coefficients, the daily RiskMetrics volatility forecast for day t is estimated as:
where \( \sigma_{t - 1}^{2} \) is the volatility of the index at day t − 1 and \( R_{t - 1}^{2} \) is the squared logarithmic return of the index on the same day.
Rights and permissions
About this article
Cite this article
Sermpinis, G., Karathanasopoulos, A., Rosillo, R. et al. Neural networks in financial trading. Ann Oper Res 297, 293–308 (2021). https://doi.org/10.1007/s10479-019-03144-y
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10479-019-03144-y