Open
Description
Feature Specs
The predict_proba method is currently a crucial part of data science within the scikit learn (sklearn) package, specifically aimed at classification problems. The method should take a data point as an input and output the probabilities that the network predicts of the input in the form [P(class 0), P(class 1), P(class 2), ...], where these probabilities add to unity. There is already a predict method on qiskit machine learning classifiers: neural_network_classifier, vqc, qsvc and pegasus_qsvc, this is to be built upon for predict_proba.
Future additions
We could also look to generalise this feature to be included on EstimatorQNN and SamplerQNN. These objects are much more flexible which makes adding the method much more of a challenge.