Both the plots show very poor results for lags. Am I missing something in terms of transformation to get proper results?
Asked
Active
Viewed 41 times
0
Darpit
- 21
- 3
-
Post your actual data. – IrishStat Sep 17 '18 at 20:39
-
@IrishStat: I have posted the data. Did you mean the actual data or just this? – Darpit Sep 17 '18 at 21:23
-
ACTUAL OBSERVED DATA – IrishStat Sep 17 '18 at 21:48
-
@IrishStat: Sir, This is the actual observed data. – Darpit Sep 17 '18 at 22:41
-
no you presented a picture of the actual data .. i want the actual data i.e. the observed values – IrishStat Sep 18 '18 at 01:17
1 Answers
0
before receipt of your actual recorded data ..
since your series has anomalies ( pulses to be sure ... ) possible seasonal pulses
since your series has possible change in error model variance over time ..
your acf is of little value as it is effected by non-gaussian structure.
correct model identification is done in a robust way.
see I have correlogram ACF and PACF below for a temperature time series. Can I say it is MA(2) from ACF? What about AR? for further comments.
IrishStat
- 27,906
- 5
- 29
- 55