Hi.
I’m looking to work with ARIMA.
But DECOMPOSE SIGNAL returns this error: lenght of values does not match lenght of index.
Welcome to the KNIME community.
What I expect is that your timeseries dataset does not have enough observations. I personaly like the book by Rob J Hyndman and George Athanasopoulos very much when looking at forecasting timeseries. It explains not only their R package fpp3, but also theory and practice.
This is taken from this book:
Some textbooks provide rules-of-thumb giving minimum sample sizes for various time series models. These are misleading and unsubstantiated in theory or practice. Further, they ignore the underlying variability of the data and often overlook the number of parameters to be estimated as well. There is, for example, no justification for the magic number of 30 often given as a minimum for ARIMA modelling. The only theoretical limit is that we need more observations than there are parameters in our forecasting model. However, in practice, we usually need substantially more observations than that.
While using different R packages on timeseries, and it’s functions in it, I noticed that some of these complain when there are less than even 36 observations.
If you want to read more on this topic you could look at Rob J Hyndman - Fitting models to short time series and the linked paper in it.
2 Likes
This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.