Leveraging the potential of 4th, 6th and 8th order Cyclci Cumulants for effective modulation classification.

This data comprises .mat matrices used in the process of AMC using deep learning classifiers. It is composed of two raw I\/Q matrices generated in matlab, the Cyclic Cumulants computedd from them and the algorithm used for their extraction. The I\/Q samples contain eight sets of 50000 signals representing 8 different digital modulations types generated with the same parameters. Carrier Frequency Offset (CFO) is the main difference between the two samples. CFO has been considered in the computation of one sample and have been neglected in the computation\/generation of the other. The modulation types include BPSK, MSK, QPSK, 16QAM\uff0c 8PSK\uff0c 64QAM\uff0c 256QAM and \u03c0\/4-DQPSK.