ml_in_rust

A lightweight, from-scratch machine learning toolkit in Rust.

Features

• Linear Algebra: Vector struct for basic operations (dot, scale, add, sub, zeros).

• Models:

  • Linear Regression (batch gradient descent, feature normalization)
  • Logistic Regression (sigmoid, cross-entropy, classification)

• Preprocessing:

  • one_hot_encoding: convert categorical columns to binary matrix
  • impute: mean imputation for missing numeric values
  • scale: standard (zero-mean/unit-variance) and min-max scaling
  • split: train/test and k-fold splitting of indices

• CSV Loading: parse CSV files into mixed-type map or numeric-only map

• Demos:

  • linear_regression_demo
  • logistic_regression_demo

Installation

# Clone the repository
git clone [email protected]:moh-dev-stack/ml_in_rust.git
cd ml_in_rust
# Build the library and demos
cargo build --release

Usage

As a library

Add to your Cargo.toml:

[dependencies]
ml_in_rust = { git = "https://github.com/moh-dev-stack/ml_in_rust" }

Then in your code:

use ml_in_rust::algebra::vector::Vector;
use ml_in_rust::models::linear_regression::gradient_descent;
// ...

Running the demos

# Linear regression demo
cargo run --bin linear_regression_demo

# Logistic regression demo
cargo run --bin logistic_regression_demo

Project Structure

src/
├─ algebra/           # Vector and linear algebra primitives
├─ models/            # Machine learning algorithms
│  ├─ linear_regression.rs
│  └─ logistic_regression.rs
├─ preprocessing/     # Feature-engineering utilities
│  ├─ one_hot_encoding.rs
│  ├─ impute.rs
│  ├─ scale.rs
│  └─ split.rs
├─ utils/             # CSV loading and parsing
└─ bin/               # Example binaries
   ├─ linear_regression_demo.rs
   └─ logistic_regression_demo.rs

tests/                # Integration tests for each module

Testing

Run the full test suite with:

cargo test

License

This project is open source under the MIT License. See LICENSE for details.