algomake 0.1.2

A custom machine learning library implemented from scratch.

AlgoMake

A custom machine learning library built from scratch to deepen understanding of fundamental algorithms.

Overview

AlgoMake is a personal project aimed at implementing various machine learning models and utility functions without relying heavily on established libraries like Scikit-learn or TensorFlow for core algorithm logic. The goal is to provide transparent and educational implementations of common machine learning concepts.

Features

• Base Estimator: A foundational BaseEstimator class for consistent API design.
• Gaussian Mixture Models (GMM): A complete implementation of GMM using the Expectation-Maximization (EM) algorithm, optimized to run without scipy dependencies for its core components.
• K-Nearest Neighbors (KNN): (Mention if implemented)
• Linear Models: (Mention if implemented, e.g., Linear Regression, Logistic Regression)
• Preprocessing Tools: (Mention if implemented, e.g., StandardScaler, MinMaxScaler)
• Evaluation Metrics: (Mention if implemented, e.g., Accuracy, MSE)
• Modular Design: Easy to extend with new algorithms.
• Comprehensive Testing: Each component is rigorously tested using pytest.

Installation

You can install AlgoMake directly from source:

git clone [https://github.com/yourusername/AlgoMake.git](https://github.com/yourusername/AlgoMake.git)
cd AlgoMake
pip install -e .