Abstract

On the basis of characteristic analysis of GM (1, 1) prediction model and Markov chain, we adopt the Grey prediction model to forecast the support cost of engineering equipment. It is suitable for time series prediction. Firstly, we improve relevant parameters of the model on the foundation of traditional GM (1, 1) model. Secondly, we map the improved gray prediction model into BP neural network with strong nonlinear processing ability and raise the nonlinear treatment processing ability of the model. Thirdly, we use the Markov chain method to amend the prediction result. The simulation result indicates that compared with traditional GM (1, 1) model, though the laminating degree to known data is not high in the model, the prediction precision improves a lot.

Key words: GM (1, 1), Grey neural network, Markov's chain, support cost.