DOI: 10.7763/IJCTE.2017.V9.1174
Dominant Eigen Vector Based Feature Selection Using Singular Value Decomposition in Automatic Modulation Classification
Abstract—Abstract—Feature based methods for Automatic Modulation Classification (AMC) have been a widely studied topic. This paper strives to design a methodology for feature selection using t-test statistics and Singular Value Decomposition (SVD) based dominant Eigen vectors. It then investigates the performance of K Nearest Neighbor (KNN) and Multiclass OnevsAll (OVA) Support Vector Machine (SVM) using the selected feature vector. Features are generated using Ambiguity Function (AF) of the modulated signals. Extensive cross validation is done to check the feature selection algorithm. Results show that Multiclass SVM classifier gives slightly better performance than KNN classifier.
Index Terms—Index Terms—Automatic modulation classification, feature extraction, singular value decomposition, classifier performance, support vector machine, k nearest neighbor, eigen vectors.
Afan Ali and Fan Yangyu are with School of Electronics and Information, Northwestern Polytechnical University, Xi’an, China (e-mail: afanali85@yahoo.com, fan_yangyu@nwpu.edu.cn).
Cite:Afan Ali and Fan Yangyu, "Dominant Eigen Vector Based Feature Selection Using Singular Value Decomposition in Automatic Modulation Classification," International Journal of Computer Theory and Engineering vol. 9, no.5, pp. 398-401, 2017.