Multiclass least-squares support vector machines for analog modulation classification

Authors:
Abdulkadir Sengur
Affiliations:
Firat University, Technical Education Faculty, Department of Electronics and Computer Science, 23119 Elazig, Turkey
Venue:
Expert Systems with Applications: An International Journal
Year:
2009

Citing 5
Cited 3

Automatic modulation recognition using time domain parameters

Signal Processing
A general approach to the automatic classification of radiocommunication signals

Signal Processing
Automatic Modulation Recognition of Communication Signals

Automatic Modulation Recognition of Communication Signals
Comparison of clustering algorithms for analog modulation classification

Expert Systems with Applications: An International Journal
Intelligent control of signal processing algorithms in communications

IEEE Journal on Selected Areas in Communications

Clustering technique-based least square support vector machine for EEG signal classification

Computer Methods and Programs in Biomedicine
Identification of motor imagery tasks through CC-LR algorithm in brain computer interface

International Journal of Bioinformatics Research and Applications
Modified CC-LR algorithm with three diverse feature sets for motor imagery tasks classification in EEG based brain-computer interface

Computer Methods and Programs in Biomedicine

Quantified Score

Hi-index	12.05

Visualization

Abstract

This study introduces the usage of multiclass least-squares support vector machines (MC-LS-SVM) for classification purposes of the analog modulated communication signals. Fulfilled study uses our previous papers where ANN and clustering methods were used as classifiers and several key features which were extracted from the instantaneous properties of the intercepted signal for characterizing the modulation types. k-fold cross-validation test, classification accuracy and confusion matrix methods are used for calculating the performance of the MC-LS-SVM classifier. Moreover, the performance of the MC-LS-SVM is compared with our previous studies where ANN and clustering efforts for modulation classification were investigated. According to the computer simulations, 100% correct classification rate was obtained when 10-fold cross-validation test method was used.