AutoML forecast only supports integers as predicate target ?

[jkibiki]

Hi Community,

I've playing around with AutoML and started with a simple forecast for Databricks samples.

I used a copy of table samples.tpch.orders.

To my supprise only integer types were available as Predicat Target. 

The field I was interested in forecasting, i.e o_totalprice ~ revenue was was listed and of type decimal.

I did not find any reference that predicate target field should be an int.

[jamesl]

Hi @jkibiki ,

It looks like decimal type is supported in 11.3 LTS ML and above -- https://docs.databricks.com/en/machine-learning/automl/automl-data-preparation.html#supported-data-f...

Let me know if that doesn't resolve your issue.

-James

[james598henry]

---

@jkibiki wrote:

Hi Community,

I've playing around with AutoML and started with a simple forecast for Databricks samples.

I used a copy of table samples.tpch.orders.

To my supprise only integer types were available as Predicat Target. 

The field I was interested in forecasting, i.e o_totalprice ~ revenue was was listed and of type decimal.

I did not find any reference that predicate target field should be an int.

---

Hello,
It seems that AutoML in Databricks currently supports only integer types as prediction targets for forecasting. This limitation might be why you're unable to use the o_totalprice field directly.

You might consider converting your target field to an integer format if possible, or exploring other machine learning tools that support decimal types for forecasting.

Best Regards,
James Henry