舰船批量建造费用的组合预测应用研究

Application research on combined forecasting of the outlay of shipbuilding in batch

研究了舰船批量建造费用的组合预测问题。首先依据舰船批量建造费用的变化特点,采用生长曲线法、数据平滑法、灰色分析法对费用进行预测,然后运用支持向量机的回归算法对3种预测方法的预测值进行有机组合,建立了舰船批量建造费用的组合预测模型,最后利用样本数据分析了一个参数变化对预测性能的影响,给出了支持向量机参数选择的依据。仿真结果表明,该模型能明显提高系统预测的精度和稳定性,可用于舰船批量建造费用的预测和估算,具有较大的实用价值。

The outlay of shipbuilding shipbuilding in batch, logistic curve, data in batch is studied. According to the special characteristic of smoothness and grey forecasting are respectively used to do forecasting. Then a combined forecasting model is established on the basis of the application of support vector regression, which is used to combine the former three methods＇ results. Finally, by making use of the data,a principle for choosing SVM parameters is given by analyzing the forecasting effect when a parameter changes. Simulation result shows that this combined model can improve both precision and stability obviously,and also can be used to forcast and estimate the outlay of shipbuilding in batch. The model has good value in utility.