Possible reasons for no prediction


I was running a sPLS-DA model and when I tried to use that to predict based on test dataset, some of the predictions were empty. The predicted values were just like this.

The way I built the model could work on some other similar datasets in my project. I have checked this test dataset, and there is no missing data at all. Is this normal? Anyone have experienced similar errors before or can give me a hint on possible reasons for this?

Thanks so much for your help.


Hi @Kathleen,

Thanks for using mixOmics.

What are the classes in Y in your sPLS-DA model? You can check using:
I’m wondering if it has actually got empty characters as class due to some sort of NA conversion, in which they will need to change.

Otherwise, I cannot reproduce this behaviour using our datasets. Can you please either provide one here, or send us an email with the code and the data.

You can click on this text to send us an email.
Alternatively, you can right-click on the above text and choose ‘Copy Email Address’.