This is my question:
I forecast the number of ticket and our revenue for our 20 point of sales .First, We put all this time series in the same strucutre and we mark revenue and number of ticket as predict. So MTS will try to forecast this value and the correlation between
this time series will be considered . This time we think that's the MTS must evaluate each time series separly so we put the measures predict only ( We think that there is no relationship between revenue on Tunis and revenue in Sfax ,Two tunisian region
We fund that the first structure with Revenue and ticket marked as predict . Is more accruate than the other murked as predict only.
Can any one help us to understand this behaviour ?
View Complete Post