Oliveira, Gustavo HFM, Leandro L. Minku, and Adriano LI Oliveira. "GMM-VRD: A Gaussian Mixture Model for Dealing With Virtual and Real Concept Drifts." 2019 International Joint Conference on Neural Networks (IJCNN). IEEE, 2019.
Most approaches focus only on real drift and fail to properly handle virtual drift.
- Proposed GMM-VRD to explicitly handle both virtual and real concept drift.
- Used probabilistic inference to decide between updating and resetting models.
- Designed distinct strategies for each drift type.
- Achieved best average accuracy compared to existing methods.
- Reduced performance degradation during drift.
# Cloning the repository
git clone https://github.com/GustavoHFMO/GMM-VRD.git
# Acessing the repository
cd GMM-VRD
# Installing the dependencies
pip install -r requirements.txt
The module GMM_batch.py shows how to train a GMM for classification using a batch of observations, and also plots the generated model.
# Running the code
python GMM_batch.py
The module GMM_online.py executes the algorithms described below in real and synthetic datasets.
# Running the code
python GMM_online.py
Oliveira, Gustavo HFM, Leandro L. Minku, and Adriano LI Oliveira. "GMM-VRD: A Gaussian Mixture Model for Dealing With Virtual and Real Concept Drifts." 2019 International Joint Conference on Neural Networks (IJCNN). IEEE, 2019.
P. R. Almeida, L. S. Oliveira, A. S. Britto Jr, and R. Sabourin, “Adapting dynamic classifier selection for concept drift,” Expert Systems with Applications, vol. 104, pp. 67–85, 2018.
L. S. Oliveira and G. E. Batista, “Igmm-cd: a gaussian mixture classification algorithm for data streams with concept drifts,” in BRACIS, 2015 Brazilian Conference on. IEEE, 2015, pp. 55–61
This project is under a GNU General Public License (GPL) Version 3. See LICENSE for more information.