A Spacecraft Electrical Characteristics Multi-Label Classification Method Based on Off-Line FCM Clustering and On-Line WPSVM
Fig 5
The steps of PCA feature extraction method have several sections: first, there are electrical data matrix X as input. Second, through the data dimension reduction, then output the principal components. Third, entering the principal components into online system, then output the result. PCA is a data dimension reduction process based on calculating the statistical covariance matrix of the spacecraft’s electronic load electrical properties. Its purpose is to find those elements which contribute most to the features of the data. The little changed elements can be got rid of, thus reducing the dimensionality so that the amount of calculation can be reduced.