Autocorrelation and Partial Autocorrelation are both statistical tools used in time series analysis, but they differ significantly. Autocorrelation measures the linear relationship between lagged values of a time series. It provides correlation at all lags, thus it can show indirect correlations.
On the other hand, Partial Autocorrelation is a measure of direct influence of previous observations on the current observation, eliminating the effect of intermediate observations. It isolates the correlation at each specific lag by removing the effects of shorter lags.
