ZAIScikit 0.1.2

A Machine learning library

Z-MachineLearningLibrary

Our Personal Machine Learning Library

This is a Machine Learning library that Abderrahmane Baidoune, Imane Rahali I would like to build in order to further our understanding of the algorithms and implement them from scratch with the help of numpy, no more.

Thus, optimization is not a concern to us, nor is documentation or code readability. Having said that, we will, and have tried to devote some effort to it !

Features

• Machine Learning Algorithms: Implementations of various machine learning algorithms.
• No Dependencies: Relies solely on numpy, avoiding other external dependencies.
• Educational Focus: Emphasis on understanding the underlying mechanics of algorithms.

Requirements

• Python 3.6 or higher
• numpy

Installation

You can install Z-MachineLearningLibrary via pip:

pip install ZAIScikit

Or clone the repository and install it locally:

git clone https://github.com/yourusername/ZAIScikit.git
cd ZAIScikit
pip install -e .

Future Improvements and additions

All Models' features need to be 2D Arrays even if there is one feature

• Better OOP Design and Redundancy Omitting
• To be implemented :
  • DBSCAN and HDBSCAN
  • UMAP
  • Reinforcement Learning
  • AlphaZero
  • Factorization Methods
  • Convolutional Neural Networks
  • RNN + LSTM
  • Transformers
• Needs Better Implementations :
  • Faster BallTree / KDTree Algorithms for KNN

Performance Comparison

Below are some comparisons of our implementations with scikit-learn's implementations.

Decision Tree Regressor Performance

Gaussian NB Performance

KMeans Clustering Performance