Kalman Filtering for Learning with Evolving Data

The source code implemented to run the experiments shown in the article is organized as follows:

• concept_drift_detector : it contains the concept drift detectors that are not present in the scikit-multiflow library
  • Adwin
  • ECDD
  • RDDM
  • STEPD
• dataset : it contains the datasets used in the experimental evaluation
  • spam
  • usent
  • usenet1
  • usenet2
  • elist (not included) available here
• evaluation : it contains the code of the k-fold distributed cross validation with the prequential evaluation mode
• models : it contains the models developed and used in the experimental evaluation
  • DriftDetectionMethodClassifier
  • K-Adwin classifier
  • KalmanNB classifier
• rewritten_code : it contains some python files from the scikit-multiflow library that have been adjusted in order to include KalmanNB at the leaves of the tree-based classifier
• testing : it contains the tests run for the experimental evaluation
  • kalmannb: it contains the tests with synthetic and real datasets for kalmannb
  • tree: it contains the tests with artificial and real datasets for tree-based models

Citing KalmanNB and HoeffdingKalmanTree

If KalmanNB and HoeffdingKalmanTree have been useful for your research and you would like to cite them in a academic publication, please use the following Bibtex entry:

@INPROCEEDINGS{ziffer2021kalman,
  author={Ziffer, Giacomo and Bernardo, Alessio and Valle, Emanuele Della and Bifet, Albert},
  booktitle={2021 IEEE International Conference on Big Data (Big Data)}, 
  title={Kalman Filtering for Learning with Evolving Data Streams}, 
  year={2021},
  organization={IEEE},
  volume={},
  number={},
  pages={5337-5346},
  doi={10.1109/BigData52589.2021.9671365}
}