The Fusion Technology Between Grey System Theories And Neural Networks And Its Application In Prediction For Time Sequence

With improving of the ability about application of IMS (information managing system), the researchers and decision-makers hold large numbers of data, and expect using math model to analysis those data and mining useful information for the sake of quantificational forecast .In the paper, aiming at one-dimension equal time interval data rows, the various amalgamation models are put forward between grey theories and artificial neural networks (ANN) owing to those extensive application area. After studying those amalgamation models, it is brought forward that the technology which is based on amalgamation between grey theories and ANN is used to mining time sequence data. In the second chapter, it is expounded about strongpoint and shortcoming of grey theories. According to the limitation of GM(1,1) and the characteristic of different data rows, the BP ANN model which based on degressive grads arithmetic of having momentum and auto-adaption function is presented ,and the model is applied to forecast the steel output of the city Pan Zhihua. The third chapter is the most important of all chapters. In the third chapter, after choosing respective strongpoint of the grey model and the ANN model ,the various models are studied which is fused the grey model and the ANN, in addition, those amalgamation models are classified as follows:1. feeble amalgamation (simple amalgamation)2. powerful amalgamation:2.1 Combination grey neural network (CGNN):2.1.1 CGNN â… 2.1.2 CGNN â…¡2.2 Full grey neural network (FGNN): 2.2.1 FGNN â… ...