MetaPerceptron (Metaheuristic-optimized Multi-Layer Perceptron) is a Python library that implements variants and the traditional version of Multi-Layer Perceptron models. These include Metaheuristic-optimized MLP models (GA, PSO, WOA, TLO, DE, ...) and Gradient Descent-optimized MLP models (SGD, Adam, Adelta, Adagrad, ...). It provides a comprehensive list of optimizers for training MLP models and is also compatible with the Scikit-Learn library. With MetaPerceptron, you can perform searches and hyperparameter tuning using the features provided by the Scikit-Learn library.

• Free software: GNU General Public License (GPL) V3 license
• Provided Estimator: MlpRegressor, MlpClassifier, MhaMlpRegressor, MhaMlpClassifier
• Provided Utility: MhaMlpTuner and MhaMlpComparator
• Total Metaheuristic-trained MLP Regressor: > 200 Models
• Total Metaheuristic-trained MLP Classifier: > 200 Models
• Total Gradient Descent-trained MLP Regressor: 12 Models
• Total Gradient Descent-trained MLP Classifier: 12 Models
• Supported performance metrics: >= 67 (47 regressions and 20 classifications)
• Documentation: https://metaperceptron.readthedocs.io
• Python versions: >= 3.8.x
• Dependencies: numpy, scipy, scikit-learn, pytorch, mealpy, pandas, permetrics.

If you want to understand how Metaheuristic is applied to Multi-Layer Perceptron, you need to read the paper titled "Let a biogeography-based optimizer train your Multi-Layer Perceptron". The paper can be accessed at the following link

Please include these citations if you plan to use this library:

@software{nguyen_van_thieu_2023_10251022,
  author       = {Nguyen Van Thieu},
  title        = {MetaPerceptron: A Standardized Framework for Metaheuristic-Trained Multi-Layer Perceptron},
  month        = dec,
  year         = 2023,
  publisher    = {Zenodo},
  doi          = {10.5281/zenodo.10251021},
  url          = {https://github.com/thieu1995/MetaPerceptron}
}

@article{van2023mealpy,
  title={MEALPY: An open-source library for latest meta-heuristic algorithms in Python},
  author={Van Thieu, Nguyen and Mirjalili, Seyedali},
  journal={Journal of Systems Architecture},
  year={2023},
  publisher={Elsevier},
  doi={10.1016/j.sysarc.2023.102871}
}

@article{van2023groundwater,
  title={Groundwater level modeling using Augmented Artificial Ecosystem Optimization},
  author={Van Thieu, Nguyen and Barma, Surajit Deb and Van Lam, To and Kisi, Ozgur and Mahesha, Amai},
  journal={Journal of Hydrology},
  volume={617},
  pages={129034},
  year={2023},
  publisher={Elsevier}
}

@article{thieu2019efficient,
  title={Efficient time-series forecasting using neural network and opposition-based coral reefs optimization},
  author={Thieu Nguyen, Tu Nguyen and Nguyen, Binh Minh and Nguyen, Giang},
  journal={International Journal of Computational Intelligence Systems},
  volume={12},
  number={2},
  pages={1144--1161},
  year={2019}
}

• Install the current PyPI release:

$ pip install metaperceptron==2.0.0

• Check the version:

$ python
>>> import metaperceptron
>>> metaperceptron.__version__

• Here is how you can import all provided classes from MetaPerceptron

from metaperceptron import DataTransformer, Data
from metaperceptron import MhaMlpRegressor, MhaMlpClassifier, MlpRegressor, MlpClassifier
from metaperceptron import MhaMlpTuner, MhaMlpComparator

• In this tutorial, we will use Genetic Algorithm to train Multi-Layer Perceptron network for classification task. For more complex examples and use cases, please check the folder examples.

from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
from metaperceptron import DataTransformer, MhaMlpClassifier

## Load the dataset
X, y = load_iris(return_X_y=True)

## Split train and test
X_train, y_train, X_test, y_test = train_test_split(X, y, test_size=0.2)

## Scale dataset with two methods: standard and minmax
dt = DataTransformer(scaling_methods=("standard", "minmax"))
X_train_scaled = dt.fit_transform(X_train)
X_test_scaled = dt.transform(X_test)

## Define Genetic Algorithm-trained Multi-Layer Perceptron
opt_paras = {"epoch": 100, "pop_size": 20}
model = MhaMlpClassifier(hidden_layers=(50, 15), act_names="Tanh", dropout_rates=None, act_output=None,
                         optim="BaseGA", optim_paras=opt_paras, obj_name="F1S", seed=42, verbose=True)
## Train the model
model.fit(X=X_train_scaled, y=y_train)

## Test the model
y_pred = model.predict(X_test)
print(y_pred)

## Print the score
print(model.score(X_test_scaled, y_test))

## Calculate some metrics
print(model.evaluate(y_true=y_test, y_pred=y_pred, list_metrics=["AS", "PS", "RS", "F2S", "CKS", "FBS"]))

• Official source code repo: link
• Official document: link
• Download releases: link
• Issue tracker: link
• Notable changes log: link
• Official chat group: link