This time we’re going bigger than ever. Fabric, Power BI, SQL, AI and more. We're covering it all. You won't want to miss it.
Learn moreDid you hear? There's a new SQL AI Developer certification (DP-800). Start preparing now and be one of the first to get certified. Register now
Hi, guys,
Is it possible to use DAX to classify datasets?
Now I need to classify customers according to their purchase characteristics, but we can take a simple dataset as an example, such as the famous iris dataset .
I need to use the old dataset to train the model and then use it to predict the new dataset, but I got stuck in clustering data sets, any ideas or relevant document ?
Solved! Go to Solution.
Hi, @DAKKK ,
I do not recommend using DAX to train algorithm models, because this is not the original intention of DAX design. I recommend that you use SQL server machine learning or sklearn to train algorithm models, and then import the prediction results into PBI.
However, if only for small data sets such as iris, you can use DAX to build a simple KNN classification model (refer to this blog),
for SQL Server Machine Learning, refer to this document.
Mark this answer as solution if it helps, thanks!
Hi, @DAKKK ,
I do not recommend using DAX to train algorithm models, because this is not the original intention of DAX design. I recommend that you use SQL server machine learning or sklearn to train algorithm models, and then import the prediction results into PBI.
However, if only for small data sets such as iris, you can use DAX to build a simple KNN classification model (refer to this blog),
for SQL Server Machine Learning, refer to this document.
Mark this answer as solution if it helps, thanks!
Check out the April 2026 Power BI update to learn about new features.
Sign up to receive a private message when registration opens and key events begin.
If you have recently started exploring Fabric, we'd love to hear how it's going. Your feedback can help with product improvements.
| User | Count |
|---|---|
| 6 | |
| 5 | |
| 5 | |
| 3 | |
| 1 |
| User | Count |
|---|---|
| 14 | |
| 7 | |
| 6 | |
| 6 | |
| 5 |