Title :
Research on serial grey neural network mode
Author_Institution :
Coll. of Mech. Eng., Heilongjiang Inst. of Sci. & Technol., Harbin, China
Abstract :
In view of the traditional grey model adaptive and self learning ability is poor, long-term prediction accuracy is not the ideal situation, on the basis of three precise grey models, BP neural network is introduced, construct the serial grey neural network model, the result of the simulation shows that, the grey neural network model prediction error and relative error does not appear increased gradually, the prediction accuracy is significantly better than the grey model and it is suitable for application.
Keywords :
adaptive systems; backpropagation; grey systems; neural nets; unsupervised learning; BP neural network; grey model adaptive ability; long-term prediction accuracy; prediction error; relative error; self learning ability; serial grey neural network mode; Adaptation models; Biological neural networks; Data models; Predictive models; Training; Vectors; grey neural network; grey prediction; model precision; neural network;
Conference_Titel :
Measurement, Information and Control (MIC), 2012 International Conference on
Conference_Location :
Harbin
Print_ISBN :
978-1-4577-1601-0
DOI :
10.1109/MIC.2012.6273440