AxML

AxML is a header-only C++20 machine learning library designed with a Scikit-Learn style API (fit, predict, transform).

Features

• Header-Only: Simple integration—just include the headers in your C++ project.
• Modern C++20: Built using modern C++ features and standard practices.
• Scikit-Learn Style Interface: Intuitive parameter configuration and consistent model workflow.

Available Algorithms & Modules

• Classification: Decision Tree, Random Forest, Logistic Regression, Support Vector Classifiers (Linear / RBF), Linear & Quadratic Discriminant Analysis (LDA / QDA), Naive Bayes.
• Regression: Linear Regression, Ridge, Lasso, ElasticNet, Decision Tree & Random Forest Regression.
• Unsupervised & Transformers: K-Means Clustering, PCA, Isolation Forest, Standard & Min-Max Scalers.
• Evaluation Metrics: Accuracy, Precision, Recall, F1 Score, MSE, MAE, R², Confusion Matrix.

Requirements

• C++20 compatible compiler (GCC 10+, Clang 10+, MSVC 2019+)
• CMake 3.20+ (for building tests/dev target)

Quick Start

Since AxML is header-only, include AxML and its deps directory in your project's include paths.

#include "AxML/AxML.hpp"
#include <iostream>

int main() {
    // 1. Configure hyperparameters
    AxML::DecisionTreeParams params;
    params.max_depth = 5;
    params.random_state = 42;

    // 2. Instantiate model
    AxML::DecisionTreeClassifier model(params);

    // 3. Fit model on feature matrix X and target y
    // model.fit(X, y);

    // 4. Predict on test data
    // AxML::VectorI preds = model.predict(X_test);

    return 0;
}

Building Tests

mkdir build && cd build
cmake ..
cmake --build .

Project Structure

• AxML/: Core library headers and algorithm implementations.
• deps/: External header dependencies.
• test/: Unit tests, data generators, and boundary evaluation scripts.

License

TBD