solarspatialtools.signalproc.correlation

solarspatialtools.signalproc.correlation(baseline, estimation, scaling='coeff')

Compute the cross correlation between two signals, including the full range of possible time lags.

The value of correlation at lags==0 is the traditional cross-correlation without allowing for time shifting.

This function is essentially a wrapper for scipy.stats.correlate, providing pre-baked scaling options, and always returning the lags.

Parameters

baselinenumeric or pandas.Series

The baseline signal

estimationnumeric or pandas.Series

The predicted signal

scalingstr

a string for the type scaling to use.

Options are:

energy:: scales by the energy of the autocorrelation of the input
coeff:: scales the output to the correlation coefficient (always removes the mean from both signals)
unbiased_energy:: similar to energy, but normalizes based on lag to account for the fewer points used in the convolution. Removes bias towards small lags.
unbiased_coeff:: similar to coeff, but normalizes based on lag to account for the fewer points used in the convolution. Removes bias towards small lags.