Prediction gives multiple results for one observation

1

Hello,

I am building my sPLSDA model for prediction, with a response variable of 2 levels. Looking at the results from running

predict(model, data)$class

I see some subjects with multiple (i.e. 2 in my case) levels assigned, in the form of LEVEL1 slash LEVEL2.

Is there a reason behind this? Does this mean my model is telling that probabilities of the subjects being the levels are the same? What would be a good way to deal with this? Thanks a lot !

Best Regards,
Chelsea

2

Hi @chelsilarious,
I can’t really comment as I would need to look at an example of the output. Also have a look at the supplemental material in https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1005752 that explains how the prediction is performed.

Kim-Anh