A Spacecraft Electrical Characteristics Multi-Label Classification Method Based on Off-Line FCM Clustering and On-Line WPSVM
Fig 4
On-line electrical characteristics identification general flow-chart.
The method using PCA feature extraction and the WPSVM classifier to providing electrical characteristics identification in the spacecraft load system is proposed. The on-line identification process uses the first training set generated off-line and the online-generated test as input. The PCA dimensionality of these data are reduced. After obtaining the PCA feature extraction vector, the input feature vector classifier is used to provide classification. Vote treatment in the presence of multiple classifiers is used to obtain the final classification result.