军用飞机维修保障费用机器学习预测方法

Machine Study Methods for Forecast of Maintenance and Support Expense of Warcraft

介绍了军用飞机维修保障费用预测方法。确定影响费用值的元素为说明性变量,应用BP人工神经网络和支持向量回归估计(Support Vector Regression,SVR)理论模型,以及收集到的维修保障费用数据进行了训练并对某型飞机的维修保障费用进行了预测,将预测结果与实际费用值进行了对比,并将神经网络预测方法、支持向量机方法和其它预测方法进行了比较。

Methods of research to maintenance and support expense of warcraft are introduced. Based on BP （ Back Propagation） network theory of neural network theory and SVR （ Support Vector Regression） theroy,the forecast models are designed. Using the data collected, maintenance and support expense of one type of warcraft is estimated, and the results are compared with the real data. The used methods are compared with other methods.