Sebastian Raschka

Since the support is a value between 0 and 1, choosing a support greater than 1 doesn't make sense and could cause confusion. Hence, it might be good to raise...

[WIP] Experiment: adding Jonathan's feat sele files

1

Recreates the PR from #834

Add an example to the `StackingClassifier` and `StackingCVClassifier` for how to get feature importances for the meta-features and meta-classifier. Use the example from #816

For quick prototyping, it might be desirable to have an option to make rougher decision region plots. See the discussion in #801

Automatic pruning for association rules

2

It might be convenient to add some automated way for pruning useless decision rules. E.g., see the discussion at #793

Add a Keras example to the `SequentialFeatureSelector` documentation, based on #777

Currently, the `no_information_rate` function in [bootstrap_point632.py](https://github.com/rasbt/mlxtend/blob/master/mlxtend/evaluate/bootstrap_point632.py) consumes too much memory if the dataset is moderately large. Needs to be improved.

We could add an explanation to the plot_decision_regions function to plot custom class names. E.g., like this ```python from mlxtend.plotting import plot_decision_regions import matplotlib.pyplot as plt from sklearn import datasets...

Add "exact" recommendation to permutation test

1

As described in #721 > The [permutation_test] function could raise a warning if the user wants to run a number of simulations that is close (> 50% ?) to the...

Adding a `topn` parameter to the Exhaustive Feature Selector

4

When running the `ExhaustiveFeatureSelector`, it is not unusual that an extraordinary large number of results get collected (due to the many combinations to evaluate). Some users reported that this leads...