I have an apparent time-series/sequencial (supervised: multi-class classification problem) dataset with each data-point time-stamped. However, some domain intuition tells me that the data is static and no need to treat them as time-series data. I have also verified that a non-timeseries method fits (even with rigorous cross-validation and other good practices) the data well. However, I want to prove this in more rigorous/methodological way, preferably in python but R is also fine if no alternative. Is there any method/test that validate my hypothesis?
$\begingroup$
$\endgroup$
3
-
2$\begingroup$ Start with some plotting $\endgroup$ – kjetil b halvorsen♦ Jan 30 '19 at 7:33
-
$\begingroup$ Basically you want to show that your data is just white noise. No autocorrelation. $\endgroup$ – user2974951 Jan 30 '19 at 8:42
-
$\begingroup$ en.wikipedia.org/wiki/Durbin%E2%80%93Watson_statistic $\endgroup$ – whuber♦ Feb 3 '19 at 14:56
Add a comment
|
