sklearn-nature-inspired-algorithms 0.15.0

Search using nature inspired algorithms over specified parameter values for an sklearn estimator.

Nature-Inspired Algorithms for scikit-learn

Nature-inspired algorithms for hyper-parameter tuning of scikit-learn models. This package uses algorithm implementations from NiaPy.

Installation

$ pip install sklearn-nature-inspired-algorithms

To install this package on Fedora, run:

$ dnf install python3-sklearn-nature-inspired-algorithms

Usage

Usage is similar to scikit-learn's GridSearchCV. Refer to the documentation for detailed guides and examples.

from sklearn_nature_inspired_algorithms.model_selection import NatureInspiredSearchCV
from sklearn.ensemble import RandomForestClassifier

param_grid = { 
    'n_estimators': range(20, 100, 20), 
    'max_depth': range(2, 40, 2),
    'min_samples_split': range(2, 20, 2), 
    'max_features': ["auto", "sqrt", "log2"],
}

clf = RandomForestClassifier(random_state=42)

nia_search = NatureInspiredSearchCV(
    clf,
    param_grid,
    algorithm='hba', # hybrid bat algorithm
    population_size=50,
    max_n_gen=100,
    max_stagnating_gen=10,
    runs=3,
    random_state=None, # or any number for reproducible results
)

nia_search.fit(X_train, y_train)

# The best parameters are stored in nia_search.best_params_.
# You can use them to train the final model.
new_clf = RandomForestClassifier(**nia_search.best_params_, random_state=42)

You can also plot the search process with a line plot or a violin plot.

from sklearn_nature_inspired_algorithms.helpers import score_by_generation_lineplot, score_by_generation_violinplot

# The line plot shows all runs. You can choose the metric to plot: 'min', 'max', 'median', or 'mean'.
score_by_generation_lineplot(nia_search, metric='max')

# For the violin plot, specify the run to plot.
score_by_generation_violinplot(nia_search, run=0)

Jupyter notebooks with full examples are available in examples/notebooks.

Using a Custom Nature-Inspired Algorithm

If you do not want to use one of the predefined algorithm configurations, you can use any algorithm from the NiaPy collection. This gives you more control over the algorithm behavior. Refer to the NiaPy documentation and examples for usage details.

Note: Use NiaPy version 2.5.1 or later.

from niapy.algorithms.basic import GeneticAlgorithm

algorithm = GeneticAlgorithm() # When a custom algorithm is provided, random_state is ignored.
algorithm.set_parameters(NP=50, Ts=5, Mr=0.25)

nia_search = NatureInspiredSearchCV(
    clf,
    param_grid,
    algorithm=algorithm,
    population_size=50,
    max_n_gen=100,
    max_stagnating_gen=20,
    runs=3,
)

nia_search.fit(X_train, y_train)

Contributing

Detailed contribution guidelines are available in CONTRIBUTING.md.