为了研究mRMR在6种方法中的必要性，对6种方法采用不含mRMR，在相

To necessity mRMR in 6 research methods, methods using six kinds of free mRMR, 50 runs under the same parameters, the accuracy of the recognition result shown in Figure 11. Seen from FIG. 11, the accuracy of the results of all the recognition, recognition accuracy is lower than the containing mRMR to verify the superiority mRMR sensitive selection feature aspect. The results also further illustrate, MWPE multi-feature fusion method still has the highest diagnostic accuracy, the method further illustrate the advantage in terms of feature extraction.

In order to study the necessity of mRMR in 6 methods, the 6 methods are used without mRMR and run 50 times under the same parameters, the correctness recognition results are shown in Figure 11. From Figure 11, it is clear that all the recognition accuracy results are lower than the recognition accuracy containing mRMR, thus verifying the superiority of mRMR in selecting sensitive features. The results further show that the MWPE multi-feature fusion method still has the highest diagnostic accuracy, and further illustrates the advantages of the method in feature extraction.

In order to study the necessity of mrmr in the six methods, the six methods were run 50 times without mrmr under the same parameters, and the accuracy recognition results are shown in Figure 11. From Figure 11, it can be seen that all the recognition accuracy results are lower than those with mrmr, thus verifying the superiority of mrmr in selecting sensitive features. The results further show that mwpe multi feature fusion method still has the highest diagnostic accuracy, and further shows the advantages of this method in feature extraction.