Aircraft neural network classification and autoregressive moving average model fault prediction method

The invention relates to an aircraft neural network classification and autoregressive moving average model fault prediction method. Comprising a principal component analysis feature extraction module, a random forest signal classification module and an autoregression moving average prediction module. When the electric signals of the aircraft are classified and predicted (101), firstly, data acquisition (102) is performed, then the data are preprocessed (103), and feature extraction is performed on the preprocessed data through a PCA method (104) to obtain feature vectors (105). And (107) voting is completed on the obtained feature vectors through a random forest algorithm (106), each feature vector can be labeled (109), each labeled data set is predicted through an ARMA algorithm (110), a prediction result is obtained (111), and the algorithm is ended (112). 本发明涉及一种飞行器神经元网络分类与自回归滑动平均模型故障预测方法。包括：主成分分析特征提取模块，随机森林信号分类模块和自回归滑动平均预测模块。当飞行器电信号进行分类与预测时(101)，首先进行数据采集(102)，然后对数据进行预处理(103)，将完成预处理的数据通过PCA方法进行特征提取(104)，得到