Introduction

Multi-target regression is a prediction task for datasets where several numerical target variables must be predicted from the same input features.

For example, you can use one model to predict several product demand indicators, several sensor measurements, or several financial quantities for each row.

Create a multi-target regression model

From a dataset, open the Lab and select Multi-target regression. Select at least two target columns.

All target columns must be numerical. DSS uses the remaining columns according to their feature roles and preprocessing settings.

DSS creates a prediction task with a dedicated Targets tab. You can change the selected targets from this tab. Changing the target list can require DSS to update the feature roles and may reset incompatible settings.