Abstract
The problem of predicting nonlinear and nonstationary signals is complex since the physical law that controls them is unknown and it is complicated to be considered. In these cases, it is necessary to devise nonlinear models that imitate or learn the rules of behavior of the problem and can be developed based on historical data. For this reason, neural networks are useful tools to deal with this type of problem due to their nonlinearly and their capacity of generalizing. This paper aims at exploring various types of neural network architectures and study their performance with time series predictions. Predictions on two sets of data (of a very different nature) will be made using three neural networks including multilayer perceptrons, recurrent neural network and long-short term memory varying some important parameters: input neurons, epochs and the anticipation with which the predictions are made. Then, all results will be compared using standard metrics.
As a conclusion, the influence of the type of series under study is more important than the parameters considered in what concerns the performance. The management of the memory in the networks is a key to its success in the prediction of S&P 500 and electrical power time series.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Brockwell, P.J., Davis, R.A.: Introduction to Time Series and Forecasting. Springer, New York (2016). https://doi.org/10.1007/b97391
Sapankevych, N.I., Sankar, R.: Time series prediction using support vector machines: a survey. IEEE Comput. Intell. Mag. 4(2), 24–38 (2009)
Dey, H.S., Kabir, M.A., Wadud, Z., Khan, S.I., Azad, M.A.K.: Econometric modeling and forecasting of natural gas demand for power sector in Bangladesh. In: TENCON 2011, pp. 1383–1386 (2011)
Liu, X.Q., Ang, B.W., Goh, T.N., Forecasting of electricity consumption: a comparison between an econometric model and a neural network model. In: IEEE International Joint Conference on Neural Networks, pp. 1254–1259 (1991)
Hochreiter, S., Schmidhuber, J.: Long short-term memory. Neural Comput. 9(8) (1997). https://doi.org/10.1162/neco.1997.9.8.1735
Lipton, Z.C.: A critical review of recurrent neural networks for sequence learning. CoRR, abs/1506.00019 (2015)
Siegelmann, H.T., Sontag, E.D.: Turing computability with neural nets. Appl. Math. Lett. 4(6) (1991). https://doi.org/10.1016/0893-9659(91)90080-F
Graves, A., et al.: Hybrid computing using a neural network with dynamic external memory. Nature 538, 471–476 (2016)
Author information
Authors and Affiliations
Corresponding authors
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2019 Springer Nature Switzerland AG
About this paper
Cite this paper
de Arquer Rilo, J., Hussain, A., Al-Taei, M., Baker, T., Al-Jumeily, D. (2019). Dynamic Neural Network for Business and Market Analysis. In: Huang, DS., Bevilacqua, V., Premaratne, P. (eds) Intelligent Computing Theories and Application. ICIC 2019. Lecture Notes in Computer Science(), vol 11643. Springer, Cham. https://doi.org/10.1007/978-3-030-26763-6_8
Download citation
DOI: https://doi.org/10.1007/978-3-030-26763-6_8
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-26762-9
Online ISBN: 978-3-030-26763-6
eBook Packages: Computer ScienceComputer Science (R0)