基于BP-MC模型的大型机电设备备件需求预测研究

针对大型机电设备备件需求具有非线性和随机波动性的特点，建立基于马尔科夫链修正的BP神经网络预测模型．以提高模型的预测精度。通过对训练样本的学习，利用BP神经网络实现了对备件需求时间序列的滚动预测，同时得到了实测值与预测值的相对误差；在此基础上利用马尔科夫链对相对误差进行修正，有效地提高了预测结果的精度。并将该模型应用于实际预测中，结果表明该模型优于BP神经网络单项预测模型，具有精度高、科学可靠的特点，为大型机电设备备件需求预测提供了新的途径。

The forecasting research for the electromechanical equipment spare parts demand based on the BP neural network model and markov chain

According to the demand of the Electromechanical equipment spare parts which has the characteristics of nonlinear and stochastic volatility; the proposed model was conducted to improve the prediction accuracy of the model, based on the BP neural network prediction model and the Markov Chain. By studying the training sample, the BP neural network realizes the rolling forecasts of time sequence for the demand of spare parts. At the same time the relative error between measured and predicted is got. Then Markov Chain is used to analysis the relative error correction. The model effectively improves the precision of predicted results. It is applied in actual forecasting. The results show that the proposed model is superior to the BP neural network prediction model, which provides a new way to predict spare parts demand of electromechanical equipment with the characteristics of high precision, reliability and scientific nature.