Interpreting a correlogram

[Latiole]

Hi,

I read a lot of things about ACF and PACF correlogramm, but I have so difficulties to interpret them.

For example, here is a correlogram (in french, "retard" = lags)



I use it to determine the number of orders in the case of an AR process (the serie is stationary).

My questions :
- Are my following conclusion right ?

The number of orders should be 5.

Thank you in advance

[noetsi]

There are two other possibilities. One is that you have a mixed model (which you likely do since the ACF indicates multiple MA terms). Second, yuo might have seasonality. It is rare from what I have seen to have that many higher order lags - so they might reflect a seasonal factor (although I don't know any that would apply at those intervals). I have rarely seen ACF and PACF that are so close to each other in terms of signficant lags even for the higher order lags. But admitedly I am new to time series too.

Everyone has trouble interpreting ACF and PACF, even the experts.