基于ARMA模型的故障率预测方法研究

Research on failure rate forecasting method based on ARMA model

在对比多种时间序列模型的基础上,结合故障率预测的特点,得出了自回归-移动平均混合模型(简称ARMA模型)是其中最适用于使用阶段故障率预测的方法这一结论。给出了基于ARMA模型的故障率预测的基本思想、预测模型、实施步骤,并按此法对某航空公司波音飞机故障率进行了预测,说明ARMA模型适用于故障率预测,且预测结果准确性较高。

According to the characteristics of failure rates, combined with the comparative results of several time series analysis methods, the conclusion that the auto-regressive moving average（ARMA） model is the most suitable failure rate forecasting method is obtained. The basic idea, the forecasting model and steps of failure rate forecasting based on the ARMA model are discussed. The failure rates of Boeing flights in a certain airline company are forecasted based on this method. Results show that the ARMA model is suitable for failure rate forecasting, and its veracity is better than traditional methods.