Soft voting predicts the class label based on class probabilities. The sums of the predicted probabilities for each classifier areg calculated for each class (which is important in the case of multiple classes). The assigned class is then the class with the maximum probability sum or argmax(p_sum).
This is recommended for an ensemble of well-calibrated classifiers, as follows:
Well calibrated classifiers are probabilistic classifiers for which the output of the predict_proba method can be directly interpreted as a confidence level.
- From the Calibration Docs on sklearn (http://scikit-learn.org/stable/modules/calibration.html)
Our code flow is identical to our hard voting classifier except that the parameter ...
