Detection and Prediction of Spectrally Correlated Processes withNon-Parallel Support Lines

In this work, we try to make inference about the detection and prediction of spectrally correlated processes with non-parallel support lines. The Loeve bi-frequency spectrum of these processes is supported by some not-parallel lines. First, the distributions of the discrete Fourier transform (DFT) and the periodogram of these processes are explored. Then, the spectral ciphering approach is used to detect and to estimate the spectral supports. Finally, the prediction problem of such processes is considered. The validity of theoretical results is studied using simulated datasets.