Performance comparison of clustering algorithms in determining the type of analog modulations

The contact between people from the first human to the present have been very important. This contact between people is called communication. Nowadays not only the people but also the machines used should be in contact with each other. This thesis also the transportation and evaluation of analog signals used by people or machines. For better quality transmitting of analog signals, modulation process is needed. Modulation is the process of transporting the signal with another transported signal. With this study, it is aimed to process the analog signals. Amplitude Modulation (AM), Double Side Band (DSB), Upper Side Band (USB), Lower Side Band (LSB), Frequency Modulation (FM), Phase Modulation (PM) of the 6 type of analog modulation using 200 samples from each of the types of carrier for the detection of intelligent clustering method has been developed. To feature extraction stage of AM, DSB, USB, LSB, FM, and PMD of the analog modulated signals are extracted from 5. level Discrete Wavelet Transform (DWT) the norm entropy and energy properties. Using the 6 × 2 × 1200 feature vector obtained in the feature extraction stage, the results were compared with the result of K-Averages, k-Medoid and Fuzzy C Averages algorithms and the result of intelligent clustering with carrier for identification. As a result of this comparison, the most successful result has been obtained with FCM (85.75%).