Abstract
The initial connection weight and threshold value of RBM in DBN have a certain influence on the recognition effect of the network. In this paper, the MNIST database is used as the data sample, and the improved genetic algorithm is applied to optimize the initial weight and threshold value in RBM, and the programming is realized by MATLAB language. The simulation results show that the improved DBN model is better than the traditional DBN network model and the traditional genetic algorithm improved DBN model in handwritten numeral recognition.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
This work was supported by Langfang Science and Technology Research and Development Self Fin- ancing Project (2019011008); the Natural Science Foundation of Hebei Province [no. F2018408040]; the Hebei Education Funds for Youth Project [no. QN2018047].
References
Murru, N., Rossini, R.: A Bayesian approach for initialization of weights in backpropagation neural net with application to character recognition. Neuroco mputing 193, 192–205 (2016)
Hanx: Handwritten digital recognition based on GA-BP neural network. In: International Conference on Artificial Intelligence, Management Science and Electronic Commerce, pp. 5050–5052. IEEE (2011)
Zhang, X., Wu, L.: Handwritten digit recognition based on improved learning rate BP algorithm. In: 2nd International Conference on Information Engineering and Computer Scienc- e Proceedings, pp. 1–4. ICIECS (2010)
Song Xiaoru, W., Xue, G.S., et al.: Simulation study on handwritten numeral recognition based on deep neural network. Sci. Technol. Eng. 19(5), 193–196 (2019)
Savich, A.W., Moussa, M.: Resource efficient arithmetic effects on RBM neural network solution quality using MNIST. In: 2011 International Conference on Reconfigurable Computing and FPGAs (ReConFig), pp. 35–40 (2011)
Guan Baolin, B.A., Lideng, B.P.: Neural network handwritten numerical recognition based on improved genetic algorithm. Control Instr. Chem. Ind. 40, 778–784 (2012). in Chinese
Chuang, C., Chellali, R., Yin, X.: Speech emotion recognition based on improved genetic algorithm optimized BP neural network. Appl. Res. Comput. 2, 344–346 (2019)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Hong-Xia, Z., Xiu-Hua, L., Hong-Ying, Z. (2021). The Application in Handwritten Digit Recognition of Deep Belief Network Based on Improved Genetic Algorithm. In: Barolli, L., Woungang, I., Enokido, T. (eds) Advanced Information Networking and Applications. AINA 2021. Lecture Notes in Networks and Systems, vol 227. Springer, Cham. https://doi.org/10.1007/978-3-030-75078-7_2
Download citation
DOI: https://doi.org/10.1007/978-3-030-75078-7_2
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-75077-0
Online ISBN: 978-3-030-75078-7
eBook Packages: Intelligent Technologies and RoboticsIntelligent Technologies and Robotics (R0)