The Gaussian Mixture Descriptors Learner (GMDL) is a cutting-edge, lightweight classifier designed for real-world, multiclass, and online classification problems. GMDL stands out for its probabilistic approach, exceptional performance with continuous features, and robustness against overfitting. These qualities make it ideal for handling large and dynamic datasets often encountered in practical applications.
Key Features
- Probabilistic Nature: Handles continuous features effectively, offering flexibility across diverse datasets.
- Online Learning Capability: Adapts dynamically to new data, maintaining performance in evolving environments.
- Overfitting Resistance: Proven robustness in experiments with real-world datasets.
Publications
- Gaussian Mixture Descriptors Learner
B.L. FREITAS, R.M. SILVA, T.A. ALMEIDA
Knowledge-Based Systems, Elsevier, Volume 188, 1-9, 2019 - Protótipo de um Método de Classificação por Descrição Mínima
B.L. FREITAS, T.A. ALMEIDA, R.M. SILVA
Anais do XIV Encontro Nacional de Inteligência Artificial e Computacional (ENIAC’17), Uberlandia, Minas Gerais, Brazil, October, 2017
Code
The source code is publicly available at this link.