Research on Communication Individual Identification Method Based on PCA-NCA and CV-SVM
In recent years, high-dimensional data has often appeared in the fields of science and industry, such as computer vision, pattern recognition, biological information, and aerospace. Feature dimension reduction and selection are the process of reducing data from high dimensionality to low dimensionality to reveal the nature of the data. In the field of wireless communication, in view of the feature redundancy caused by the high-dimensional features of wireless device startup transient signals, this paper converts the high-dimensional features of signals into low-dimensional features that are conducive to classification through the feature dimensionality reduction and selection method based on PCA-NCA. In addition, this paper also carried out parameter optimization for SVM classifier, and the established CV-SVM classifier improved the classification performance. This paper also carries out simulation devices on the measured start-up signals of ten identical walkie-talkies. When the SNR is greater than 0 dB, the recognition accuracy of the PCA-NCA algorithm is 10% higher than recognition accuracy of the PCA algorithm alone; when the SNR is greater than 10 dB.
KeywordsPCA Individual identification Feature selection
The authors would like to thank State Key Laboratory of Complex Electromagnetic Environment Effects on Electronics and Information System Director Fund (CEMEE2019K0104B).
- 6.Martin-Bautista, M.J., Vila, M.A.: A survey of genetic feature selection in mining issues. In: Proceeding of the 1999 Congress on Evolutionary Computation, pp. 13–23. IEEE Press (1999)Google Scholar
- 7.Goldberger, J., Roweis, S., Hinton, G., et al.: Neighbourhood components analysis. In: Proceedings of 2005 Conference on Neural Information Processing Systems. MIT Press (2005)Google Scholar
- 8.Singh-Miller, N., Collins, M., Hazen, T.J.: Dimensionality reduction for speech recognition using neighborhood components analysis. In: Proceedings of Interspeech 2007, Antwerp, Belgium (2007)Google Scholar
- 10.Xing, E.P., Ng, A.Y., Jordan, M.I., et al.: Distance metric learning, with application to clustering with side-information. In: Proceedings of 2002 Conference on Neural Information Processing Systems. MIT Press, Cambridge (2002)Google Scholar
- 11.Globerson, A., Roweis, S.T.: Metric learning by collapsing classes. In: Proceedings of 2006 Conference on Neural Information Processing Systems. MIT Press, Cambridge (2006)Google Scholar