Research on interference signal recognition based on nonlinear support vector machine and logistic regression

In this paper, the extraction of interference signals in acoustic emission and electromagnetic radiation and identifying time intervals for interference signal are studied. Firstly, select 11 kinds of interference characteristics in time domain, frequency domain and wavelet domain. Then, combine the Spearman coefficient and mutual information to extract three interference features, namely mean, standard deviation and wavelet domain energy for the test data. Then, establish the logistic regression model and the nonlinear support vector machine model (SVM)with the three interference features as the input vectors. Finally, carry out numerical experiments to obtain the time intervals of the interference signal.