statsmodels.tsa.stattools.ccf¶
-
statsmodels.tsa.stattools.ccf(x, y, unbiased=True)[source]¶ cross-correlation function for 1d
- Parameters
- x, yarrays
time series data
- unbiasedboolean
if True, then denominators for autocovariance is n-k, otherwise n
- Returns
- ccfarray
cross-correlation function of x and y
Notes
This is based np.correlate which does full convolution. For very long time series it is recommended to use fft convolution instead.
If unbiased is true, the denominator for the autocovariance is adjusted but the autocorrelation is not an unbiased estimtor.