[AutoML] Add AutoML objective evaluation for classification tasks

[bcebere]

Feature Description

For a sampled pipeline, the AutoML search process needs to evaluate an objective, and suggest future points in a direction.

For classification tasks, the evaluation metrics are documented here #11
The benchmark is done using the cross-validation tester documented in #20.
Given a metric, the optimization process might seek to maximize or minimize the objective.

AP Reference: https://github.com/vanderschaarlab/autoprognosis/blob/main/src/autoprognosis/explorers/classifiers.py