Handling time granularity using the forecast recipe
lisa811
Partner, Dataiku DSS Core Designer, Dataiku DSS & SQL, Dataiku DSS ML Practitioner, Dataiku DSS Core Concepts, Registered Posts: 14 Partner
Hello,
I am trying to do a forecast using the recipe. The input I've got is sampled at hour granularity. The problem is that I do not have values for every hour. If there is no value, it should just keep the one from the row/hour before, because that means that no change has occured. But when using the recipe, I can only choose to use a fixed value or interpolate. This does not refelct my data in an appropriate way. Is there any way how to fix that?
I am trying to do a forecast using the recipe. The input I've got is sampled at hour granularity. The problem is that I do not have values for every hour. If there is no value, it should just keep the one from the row/hour before, because that means that no change has occured. But when using the recipe, I can only choose to use a fixed value or interpolate. This does not refelct my data in an appropriate way. Is there any way how to fix that?
Tagged:
Answers
-
Hi,
Right now it is possible by adding a Prepare recipe using the "Fill empty cells with previous/next value": https://doc.dataiku.com/dss/latest/preparation/processors/up-down-fill.html
I added your feature request to our backlog for future enhancements of the plugin.
Best regards,
Alex Combessie
