algomake 0.1.3

A custom machine learning library implemented from scratch.

AlgoMake

AlgoMake is a powerful Machine Learning library designed for deep understanding, with every algorithm meticulously built from scratch using pure NumPy. Explore the core mechanics of ML without abstraction.

---

Overview

AlgoMake offers a unique approach to learning and implementing machine learning. It's a comprehensive collection of fundamental algorithms, each developed from first principles with a strong emphasis on clarity and mathematical accuracy, relying solely on NumPy.

This project serves as an invaluable resource for anyone looking to gain a transparent and educational understanding of machine learning models.

---

Why Choose AlgoMake?

In a landscape dominated by high-level ML frameworks, AlgoMake stands out by providing:

• Unparalleled Transparency: Every algorithm's core logic is exposed, allowing users to see exactly how computations are performed.
• Pure NumPy Implementation: Eliminates external dependencies for core algorithm computations, making the underlying mathematics explicit.
• Enhanced Learning: Ideal for students and practitioners who want to master the mathematical foundations and inner workings of ML models.

---

Key Features

AlgoMake currently includes a growing suite of carefully engineered algorithms:

Core Algorithms (algomake/models/)

• Gaussian Mixture Models (GMM):

  • Complete Expectation-Maximization (EM) algorithm from scratch
  • Custom implementation of the multivariate Gaussian PDF
  • Robust handling of numerical stability

• Support Vector Machines (SVM):

  • Planned: Dual formulation, Hinge loss, Kernel trick, SMO algorithm

• Other Models:

  • Linear Regression, Logistic Regression
  • K-Nearest Neighbors (KNN)
  • Decision Trees
  • Ensemble Methods (Bagging, Boosting)
  • Clustering Algorithms (K-Means)

Preprocessing & Dimensionality Reduction (algomake/preprocessing/)

• Principal Component Analysis (PCA):

  • Manual computation of eigenvectors and eigenvalues
  • Covariance matrix and dimensionality reduction pipeline

• Standardization/Normalization

Utility Components

• BaseEstimator: A foundational class providing a consistent fit, predict, get_params, and set_params interface.
• Metrics: Custom implementations of classification and regression evaluation metrics.

---

Installation

To integrate AlgoMake into your Python environment:

git clone https://github.com/ShutterStack/AlgoMake.git
cd algomake
pip install -e .

For development (includes testing, formatting, etc.):

pip install -e .[dev]

---

Usage Examples

Gaussian Mixture Models (GMM)

import numpy as np
from algomake.models.gmm import GaussianMixture

# Generate sample data
np.random.seed(0)
data_1 = np.random.multivariate_normal([2.0, 2.0], [[0.5, 0.2], [0.2, 0.5]], 100)
data_2 = np.random.multivariate_normal([8.0, 8.0], [[0.7, -0.3], [-0.3, 0.7]], 100)
X_train = np.vstack((data_1, data_2))

gmm = GaussianMixture(n_components=2, random_state=42, max_iter=100, tol=1e-4)
gmm.fit(X_train)

X_new = np.array([[2.1, 1.9], [8.3, 7.8], [5.0, 5.0]])
predicted_labels = gmm.predict(X_new)
probabilities = gmm.predict_proba(X_new)

print("Labels:", predicted_labels)
print("Probabilities:", np.round(probabilities, 4))

---

Principal Component Analysis (PCA)

import numpy as np
from algomake.preprocessing.dimensionality_reduction import PCA

X = np.array([
    [2.5, 2.4], [0.5, 0.7], [2.2, 2.9],
    [1.9, 2.2], [3.1, 3.0], [2.3, 2.7],
    [2.0, 1.6], [1.0, 1.1], [1.5, 1.6], [1.1, 0.9]
])

pca = PCA(n_components=1)
pca.fit(X)
X_transformed = pca.transform(X)

print("Transformed Data:", np.round(X_transformed, 3))

---

Running Tests

Run all unit tests using pytest:

pytest

---

Contributing

We welcome and encourage contributions! 🚀

To contribute:

1. Fork the repository
2. Clone your fork:

   git clone https://github.com/ShutterStack/AlgoMake.git
   

3. Create a feature branch:

   git checkout -b feature/your-feature-name
   

4. Implement changes
5. Add unit tests
6. Run pytest to ensure all tests pass
7. Format with black and isort
8. Commit and push your changes
9. Open a Pull Request

---

License

This project is licensed under the MIT License.

---

Contact

Have questions, suggestions, or want to collaborate?

• GitHub Issues: Submit here
• Email: [e-mail]

---

Enjoy learning ML from the inside out with AlgoMake!