A simple way to graphically present the results of a TSA is to plot
the test statistics S against its parameter
or l, depending
on whether the analysis was performed in the frequency or time domain.
Plots in the frequency domain are called periodograms. In them,
oscillations are revealed by the presence of spectral lines. However,
some (often many) lines are spurious and simply arise from random
fluctuations of the signal. By means of the confidence level
and the probability distribution of S one can find the critical
value Scrit for significant features. Examples of statistics
used in the time domain are covariance and correlation functions. The
correlation of a signal with itself or with another signal produces
maxima in these functions at particular lags. Detection of genuine
lags then consists of testing the significance of such maxima.