View model's predictions and probabilities of each prediction class about transmogrifai HOT 4 CLOSED

are you running it on 0.3.4 release? if so, hen simply add raw prediction and probability into your result features:

val (pred, rawPred, prob) = modelSelector.getOutput()
val model = new OpWorkFlow()setReader(reader).setResultFeatures(pred, rawPred, prob).train()
val scores = model.score()

More on model selectors - https://docs.transmogrif.ai/en/stable/developer-guide/index.html#modelselector

from transmogrifai.

In the upcoming 0.4.0 release the result type of the model selector will be the combined type that contains all three values - http://docs.transmogrif.ai/en/master/developer-guide/#modelselector

from transmogrifai.

Many thanks for your answer!
What about RegressionModelSelector?
Probabilities are missing because the underlying models are not producing it?

from transmogrifai.

Correct.

from transmogrifai.