Hello,
I want to do a time series forecast in data that I wish to group by week and I use PIVOT to do that. The problem is that then I lose the date type and I don't want it. Before Rapidminer I use to use Excel to transform year and week numbers in date (as first day of each week). By doing this I get a monotonic time series.
In Excel I used to calculate the first day of each week like below, but I want to perform all data prep in RM and honestly I'm getting nowhere.
=DATE([
@YEAR],1,-2)-WEEKDAY(DATE([
@YEAR],1,3))+[
@WEEK]*7I bet RM can do this easily with generate attributes operator, but I'm totally stucked.
Thanks for your help,
Pedro