basicMLpy 1.0.1

A collection of simple machine learning algorithms

basicMLpy

basicMLpy is a package that implements simple machine learning algorithms. It currently contains seven modules that implement multiple machine learning techniques for supervised learning.

The basicMLpy.regression module contains the following functionalities:

• Linear Regression
• Ridge Regression

The basicMLpy.classification module contains the following functionalities:

• Multiclass classification through the IRLS(Iteratively Reweighted Least Squares) algorithm

The basicMLpy.nearest_neighbors module contains the following functionalities:

• An implementation of the K-Nearest Neighbors algorithm, that can fit both classification and regression problems

The basicMLpy.model_selection module contains the following functionalities:

• A Cross-Validation algorithm for the functions presented by the basicMLpy package

The basicMLpy.ensemble module contains the following functionalities:

• An implementation of the Random Forests algorithm for regression and classification
• An implementation of the AdaBoost algorithm for classification
• An implementation of the Gradient Boosting algorithm for regression

The basicMLpy.loss_functions module contains the following functionalities:

• Multiple functions for error evaluation, e.g. MSE, MAE, exponential loss, etc.

The basicMLpy.utils module contains the following functionalities:

• Useful functions utilized all throughout the other models.

Documentation

The documentation will be available at a proper site soon. For now it can be found in the main code for each module.

Installation

To install basicMLpy run the following command:
pip install basicMLpy

Dependencies

basicMLpy requires Python >= 3.8, Numpy >= 1.19, Scipy >= 1.5.2, scikit-learn >= 0.23.

On Github

https://github.com/HenrySilvaCS/basicMLpy

Some thoughts

This is a work in progress project, so more functionalities will be added with time. The main feature that will be implemented in the near future is a decomposition module.

Author

Henrique Soares Assumpção e Silva