ML3 – Multiclass Latent Locally Linear Support Vector Machine

A C++ implementation with a MATLAB © interface of the ML3 SVM classifier (Fornoni et al., ACML 2013), an efficient non-linear classifier based on a latent SVM formulation.

ML3 can learn complex decision functions (traditionally given by kernels) through the use of locally linear decision functions. Differently from kernel classifiers, ML3 makes use of a set of linear models that are locally linearly combined (for each sample and class) to form a non-linear decision boundary in the input space. Thanks to the latent formulation, the combination coefficients are modeled as latent variables and efficiently estimated using an analytic solution.

Quick download | Official page | Github

---

Mathematica SVM – A hands-on introduction to Support Vector Machines using Mathematica ©

This project presents the very basic theory of linear classifiers, max-margin classifiers and Support Vector Machines and explores the use of Mathematica © to solve the optimization problems that arise.

Following the presentation in [1], this notebook explicitly derives, implements and compares several classifiers, demonstrating them on synthetic 2D-data generated by the user, with visualizations involving direct hyper-parameters manipulations.
The project can be considered a hands-on introduction to the topic.

[1] Nello Cristianini and John Shawe-Taylor. An introduction to Support Vector Machines and other kernel-based learning methods. Cambridge university press, 2000

Quick download | Github

---