GM(1,1)和SVR的雷达电子部件状态趋势预测模型

为提高雷达电子部件状态趋势预测的精度,根据测试数据特点,提出了基于GM(1,1)与支持向量机回归(SVR)的组合预测模型。采用粒子群优化算法分别对GM(1,1)和SVR模型进行了改进,提高了单一模型的预测精度。在此基础上,结合GM(1,1)模型对趋向性数据的预测优势和SVR模型对数据波动的强适应性,达到了取长补短、相得益彰的效果。实验结果表明该组合模型不但具有更高的预测精度,而且对不同预测对象有更强的适应能力。

Radar Electronic Components State Trend Forecast Model Based on GM (1,1) and SVR

In order to improve the forecast accuracy of radar electronic components state trend, according to the characteristics of test data,this paper proposes a combination forecasting model based on GM(1,1) and support vector machine regression(SVR).By using PSO,GM(1,1) and SVR model are separately modified to improve the forecast accuracy.On this basis,the combined model combines with GM (1,1) model’ s advantages of the trend data forecast and SVR model’ s strong adaptability to data fluctuations,reaching each other,complement each other.Experimental results show that the combined model not only has higher forecast accuracy,but stronger adaptability to different objects.