基于BP神经网络-马尔科夫模型的BDI短期预测

Short-term Prediction of BDI Based on BP Neural Network-Markov Model

通过建立BP神经网络模型,利用相邻的BDI数据作为输入变量进行预测,避开了BDI短期预测时影响因素难以量化等问题,同时,利用马尔科夫模型修正神经网络的预测结果,提高预测精度,从而提出一种可用于BDI短期预测的方法。此研究对进出口企业正确核算商品进出口成本、航运公司合理分配运力以及提高自身竞争力等方面都具有重要意义。

By establishing BP neural network model and using adjacent BDI data as input variables to predict, the problem of difficult quantification of influencing factors in short-term prediction of BDI is avoided. At the same time, Markov model is used to modify the prediction results of the neural network and improve the prediction accuracy, thus a method for short-term prediction of BDI is proposed. This study is of great significance for import and export enterprises to correctly account for the cost of import and export of commodities, to rationally distribute the capacity of shipping companies, and to improve their competitiveness.