This document showcases some cool features of Trefle.
Since Trefle is compatible with scikit-learn you can perform a grid search
using GridSearchCV. You can see an example of that in the file
/examples/grid_search_example.py.
TrefleClassifier class has a lot of parameters you can tune. Fortunately, most
of them use default values that "are good enough" for small problems.
There are two kinds of parameters, fuzzy systems and evolutionary parameters.
These parameters are listed in TrefleClassifier class and are detailed in
the classes CocoIndividual
and CocoExperiment.
You can change the fitness function to make Trefle generate models that match your criteria. For example you might want to have models that have a high sensibility. Check /examples/grid_search_example.py out.