Accessing probabilities from a model as a trade condition

[markxbaker]

Hi,

Trying to access the proba from a randomforest classifier in my back test to use as part of the condition to enter, e.g only enter if above 0.6 but having no luck - anyone tried similar with success?

Short of training model separately and registering as a new column which defeats purpose of having all in one place, I can't think of any other any ideas?

[edtechre]

Hi @markxbaker,

Have you tried passing a custom predict_fn to model()? https://www.pybroker.com/en/latest/reference/pybroker.model.html#pybroker.model.model

[markxbaker]

Thanks Ed:)