Neural Computing and Applications

, Volume 31, Supplement 1, pp 573–582 | Cite as

Research on exchange rate forecasting based on deep belief network

  • Jing Zheng
  • Xiao FuEmail author
  • Guijun Zhang
Original Article


Exchange rate forecasting has always been a research hot spot of international finance studies. Deep belief network (DBN) model of deep learning is a new method of predicting the exchange rate data, and the designing of DBN structure and the learning rules of parameters are the most important parts of DBN model. The paper firstly divides the time series data into training and testing sets. By optimizing the DBN parameters, the paper analyses the results of the training analysis and answers how to do node setting. Then, the paper adjusts the number of hidden nodes, inputs nodes and hidden layers, and by using multiple variance analysis, it determines the sensitive range of the node. Finally, the experiments of INR/USD and CNY/USD have proved that compared with the FFNN model, the improved DBN model could better forecast the exchange rate.


Deep learning DBN Exchange rate forecasting 


Compliance with ethical standards

Conflict of interest

The authors declared that they have no conflicts of interest to this work. We declare that we do not have any commercial or associative interest that represents a conflict of interest in connection with the work submitted. This work was supported in part by the Scientific Research Starting Foundation of Hangzhou Dianzi University (No. KYS395617020). The author Fu X. also works in the Research Center of Information Technology & Economic and Social Development.


  1. 1.
    Zhang G, Hu MY (1998) Neural network forecasting of the British pound/US dollar exchange rate. Omega 26(4):495–506CrossRefGoogle Scholar
  2. 2.
    Zhang R, Shen F, Zhao J (2014) A model with fuzzy granulation and deep belief networks for exchange rate forecasting. In: International joint conference on neural networks, pp 366–373Google Scholar
  3. 3.
    LeCun Y, Bengio Y, Hinton G (2015) Deep learning. Nature 521(7553):436–444CrossRefGoogle Scholar
  4. 4.
    Schmidhuber J (2015) Deep learning in neural networks: an overview. Neural Netw 61:85–117CrossRefGoogle Scholar
  5. 5.
    Hinton GE, Osindero S, Teh YW (2006) A fast learning algorithm for deep belief nets. Neural Comput 18:1527–1554MathSciNetCrossRefzbMATHGoogle Scholar
  6. 6.
    Hinton GE, Salakhutdinov RR (2008) Using deep belief nets to learn covariance kernels for Gaussian processes. In: Advances in neural information processing systems, pp 1249–1256Google Scholar
  7. 7.
    Kuremotoa T, Kimuraa S, Kobayashib K, Obayashia M (2014) Time series forecasting using a deep belief network with restricted Boltzmann machines. Neurocomputing 137(15):47–56CrossRefGoogle Scholar
  8. 8.
    Shen F, Chao J, Zhao J (2015) Forecasting exchange rate using deep belief networks and conjugate gradient method. Neurocomputing 167(C):243–253CrossRefGoogle Scholar
  9. 9.
    Chen H, Murray AF (2003) Continuous restricted Boltzmann machine with an implementable training algorithm. IEE Proc Vis Image Signal Process 150(3):153–158CrossRefGoogle Scholar
  10. 10.
    Minsky M, Papert S (1988) Perceptrons, vol 84(1). MIT Press, Cambridge, pp 449–452zbMATHGoogle Scholar
  11. 11.
    Erhan D, Manzagol PA, Bengio Y, Bengio S, Vincent P (2009) The difficulty of training deep architectures and the effect of unsupervised pre-training. In: Immunology of fungal infections, vol 5, pp 153–160Google Scholar
  12. 12.
    Erhan D, Bengio Y, Courville A, Manzagol PA, Vincent P (2010) Why does unsupervised pre-training help deep learning? J Mach Learn Res 11:625–660MathSciNetzbMATHGoogle Scholar
  13. 13.
    Hinton GE, Dayan P, Frey BJ (1995) The “wake–sleep” algorithm for unsupervised neural networks. Science 268(5214):1158–1161CrossRefGoogle Scholar
  14. 14.
    Hinton GE (2002) Training products of experts by minimizing contrastive divergence. Neural Comput 14(8):1771–1800CrossRefzbMATHGoogle Scholar
  15. 15.
    Ma X, Wang X (2016) Average contrastive divergence for training restricted Boltzmann machines. Entropy 18(1):35CrossRefGoogle Scholar
  16. 16.
    Chen H, Murray A (2002) A continuous restricted Boltzmann machine with a hardware-amenable learning algorithm. In: Artificial neural networks—ICANN 2002, pp 358–363Google Scholar
  17. 17.
    Frey BJ (1997) Continuous sigmoidal belief networks trained using slice sampling. In: Advances in neural information processing systems, pp 452–458Google Scholar
  18. 18.
    Khashei M, Bijari M, Ardali GAR (2009) Improvement of auto-regressive integrated moving average models using fuzzy logic and artificial neural networks. Neurocomputing 72(4):956–967CrossRefGoogle Scholar
  19. 19.
    Dhamija AK, Bhalla VK (2011) Exchange rate forecasting: comparison of various architectures of neural networks. Neural Comput Appl 20(3):355–363CrossRefGoogle Scholar
  20. 20.
    Kodogiannis V, Lolis A (2002) Forecasting financial time series using neural network and fuzzy system-based techniques. Neural Comput Appl 11(2):90–102CrossRefGoogle Scholar

Copyright information

© The Natural Computing Applications Forum 2017

Authors and Affiliations

  1. 1.College of EconomicsHangzhou Dianzi UniversityHangzhouChina
  2. 2.Institute of Innovation and DevelopmentHangzhou Dianzi UniversityHangzhouChina
  3. 3.Institute of Applied MathematicsHangzhou Dianzi UniversityHangzhouChina

Personalised recommendations