White Noise Prob

[johanihanan]

I have issue with interpreting the results of residual autocorrelation  for ARIMA model.  The plot of lag versus White Noise Prob is empty! seems the values are very small. The p value is 0.003.  Based on P value , the null hypothesis of no autocorrelation in the residual is rejected at 5%. that means there are strong correlation BUT the plot shows small autocorrelation values!

I hope someone help me on that 

thank you

[Ksharp]

White Noise is special term in time series analysis. So post it at forecast forum would be better.

https://communities.sas.com/t5/SAS-Forecasting-and-Econometrics/bd-p/forecasting_econometrics

[Ksharp]

Also could check PROC AUTOREG in Rick's blog:

https://blogs.sas.com/content/iml/2015/03/12/digits-of-pi.html

[sbxkoenk · Accepted solution]

Hello,

Can you specify which two plots you are talking about?

Is it:

• RESIDUAL(WN) --> produces the plot of Ljung-Box white-noise test p-values at different lags.
• RESIDUAL(ACF) --> produces the plot of residual autocorrelations.

Or maybe you have just asked for
RESIDUAL(CORR)?

What is small?
Small autocorrelation values can also be significantly different from zero of course.

BR, Koen