I came across some very strange ACF, PACF chart during my research. how should these charts be interpreted? From top to bottom are chart of the original time series, the acf chart and the pacf chart
$\begingroup$
$\endgroup$
1
-
1$\begingroup$ Hi, I can't say for sure, but I think your ACF plot (which I'm assuming is the first one) may suggest that there is some form of apparent non-stationarity due to the very slow decay which, would in a sense, invalidate the use of acf/pacfs calculated on the whole data set in it's current state since while you can calculate the acf/pacf for various lags, it won't necessarily mean something, e.g. like calculating the sample mean of draws from a distribution with no first moment. $\endgroup$– QMathSep 13 at 3:05
Add a comment
|