Concept Drift Detection Through Resampling
Abstract
Detecting changes in data-streams is an important part of enhancing learning quality in dynamic environments. We devise a procedure for detecting concept drifts in data-streams that relies on analyzing the empirical loss of learning algorithms. Our method is based on obtaining statistics from the loss distribution by reusing the data multiple times via resampling. We present theoretical guarantees for the proposed procedure based on the stability of the underlying learning algorithms. Experimental results show that the detection method has high recall and precision, and performs well in the presence of noise.
Cite
Text
Harel et al. "Concept Drift Detection Through Resampling." International Conference on Machine Learning, 2014.Markdown
[Harel et al. "Concept Drift Detection Through Resampling." International Conference on Machine Learning, 2014.](https://mlanthology.org/icml/2014/harel2014icml-concept/)BibTeX
@inproceedings{harel2014icml-concept,
title = {{Concept Drift Detection Through Resampling}},
author = {Harel, Maayan and Mannor, Shie and El-Yaniv, Ran and Crammer, Koby},
booktitle = {International Conference on Machine Learning},
year = {2014},
pages = {1009-1017},
volume = {32},
url = {https://mlanthology.org/icml/2014/harel2014icml-concept/}
}