Hi Valeriy,

I'd say so, particularly if the time series in the dataset are not related.

Two time series can have similar dynamics, but exhibit different value ranges. I think normalization can help in these cases, but this is probably situational and can be easily tested using CV.