mlmechanica 0.1.0

A transparent, 'Glass Box' Machine Learning library designed for education.

MLMechanica

Machine Learning, Unveiled.

mlmechanica is a custom Machine Learning library built from scratch in Python. Unlike traditional libraries that treat models as "black boxes," MLMechanica is designed to be a "Glass Box"---offering complete transparency into the mathematical operations, internal states, and iterative steps of every algorithm.

It is strictly educational, optimized for clarity and understanding rather than production speed.

---

🚀 Key Features

• Transparency First: Enable the calculation=True flag to see every matrix multiplication, gradient update, and intermediate derivation logged to your console in real-time.
• Pure Python & NumPy: Implementations rely solely on NumPy for linear algebra, avoiding high-level abstractions to show exactly how the math works.
• Self-Documenting Models: Every class includes a model_analysis() method that returns the rigorous mathematical derivation and theory behind that specific algorithm.
• Instant Demos: Built-in static demo() methods allow you to run smoke tests and visualize model performance instantly without writing setup code.

---

📦 Installation

From Source

You can clone the repository directly from GitHub:

git clone https://github.com/Sarbik-Mal/mlmechanica.git
cd mlmechanica
pip install .

(Note: PyPI installation coming soon via pip install mlmechanica)

---

⚡ Quick Start

1. Run a built-in Demo

Want to see Lasso Regression in action immediately? Every model comes with a static demo that generates synthetic data, trains the model, and evaluates it.

from mlmechanica.regression.linear import LassoRegression


LassoRegression.demo()

2. Custom Usage with "Calculation Mode"

See the internal math (Gradient Descent, Matrix Inversion, etc.) by setting calculation=True.

import numpy as np
from mlmechanica.regression.linear import MultipleLinearRegression

X = np.array([[1, 2], [2, 3], [3, 4], [4, 5]])
y = np.array([2, 3, 4, 5])

model = MultipleLinearRegression(calculation=True)

model.fit(X, y)

pred = model.predict(np.array([[5, 6]]))
print(f"Prediction: {pred}")

---

📚 Supported Models

Currently, the library focuses on linear regression techniques:

Module	Class	Description
Simple Linear	SimpleLinearRegression	Univariate regression using closed-form OLS derivation.
Multiple Linear	MultipleLinearRegression	Multivariate regression using the Normal Equation (Vectorized).
Lasso	LassoRegression	L1 Regularization using Coordinate Descent and Soft Thresholding.
Ridge	RidgeRegression	L2 Regularization offering multiple solvers: lsqr, svd, cholesky, and mbsag (Stochastic Avg Gradient).

🧠 Model Analysis

Retrieve the mathematical derivation directly from any model:

from mlmechanica.regression.linear import SimpleLinearRegression

model = SimpleLinearRegression()

print(model.model_analysis('derivation'))

---

🤝 Contributing

Contributions are welcome! This is an educational project, so clarity and readability are prioritized.

1. Fork the Project\
2. Create a Feature Branch (git checkout -b feature/NewAlgorithm)\
3. Commit Changes (git commit -m 'Add DecisionTree implementation')\
4. Push to Branch (git push origin feature/NewAlgorithm)\
5. Open a Pull Request

---

📄 License

Distributed under the MIT License. See LICENSE for more information.