Bispectrum is a higher-order extension of power spectral estimation.

A conventional power spectrum decomposes the power of a time series over frequency. In contrast, the bispectrum decomposes the third moment (skewness) of a time series over frequency. You can use the estimated bispectrum to detect asymmetric nonlinearities in a time series.

Use the TSA Bispectrum VI to compute the bispectrum of a time series. You can use this VI to compute the bispectrum based on the Fast Fourier Transform (FFT) or the AR model of the time series.