优化的灰色马尔科夫外汇支出预测模型

Predict Model for Foreign Exchange Expenditure Based on Optimized Grey-Markov

摘要: 本文以灰色预测为基础，结合马尔科夫模型，建立灰色马尔科夫预测模型。然后使用粒子群算法对参数进行优化，提高预测精度。将灰色模型和马尔科夫模型组合，用灰色模型得到时间序列数据的趋势，用马尔科夫模型使数据呈现出波动性，得到时间序列趋势预测。通过对外汇支出业务量的数据进行验证，结果表明：粒子群优化的组合预测模型即可得到时间序列数据的总体趋势，又可以使数据随着时间的变换呈现出变换的波动性，提高预测精度。优化的预测组合预测模型精度高于单一预测模型的精度。

Abstract: Based on the grey prediction, and Markov model, the Grey-Markov prediction model is proposed in the paper. Then PSO is used to optimize the parameters and improve the prediction accuracy. Combining grey model and Markov model, the trend of time series data is obtained by grey model. Markov model makes the data show volatility, to obtain the time series trend prediction. By verifying the data on the foreign exchange expenditure business volume, the results show that the combined prediction model of PSO can not only get the general trend of time series, but also make the data show the volatility with time, and improve the prediction accuracy. The accuracy of the optimized combined forecasting model is higher than that of the single forecasting model.