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:
MhaMlpTunerandMhaMlpComparator - 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"]))