Error with covariate in ARIMA TS Factory

Question

Hi,

I have a group of time series that I want to forecast. I'm using the WF here below which is pretty straight forward. I have 4 fields :

- Grouping field

- Count (value I want to forecast)

- Date

- Covariate (0 / 0.5 / 1). This covariate is an attribute of the group ; it has the same value for the whole time series of each group.

I get the following errors when I run my WF :

TS Model Factory (4) Tool #64: Error in optim(init[mask], armaCSS, method = optim.method, hessian = FALSE, :
TS Model Factory (4) The matrix of covariates is not of full rank. Thus, the covariates for model xxxxxxxxx will not be used in model estimation.
TS Model Factory (4) Take caution when using this model as input into the TS Forecast Factory tool.

Should I prepare differently my dataset ?

Thanks.

TonyM · Accepted Answer

Hi @gmerce,

The error is coming from the covariate being a constant in the equation. Since the covariate you are trying to use is the same within each group you will get the rank deficiency error. When using covariates you will need some kind of variance within that variable in order to run it through an ARIMA (or any regression model).

A matrix is said to have full rank if its rank equals the largest possible for a matrix of the same dimensions, which is the lesser of the number of rows and columns. A matrix is said to be rank deficient if it does not have full rank.

https://en.wikipedia.org/wiki/Rank_(linear_algebra)

Hope this helps.