Springe direkt zu Inhalt

Active feature acquisition on data streams under feature drift

Eirini Ntoutsi, Christian Beyer, Maik Büttner, Vishnu Unnikrishnan, Miro Schleicher, Eirini Ntoutsi, Myra Spiliopoulou – 2020

Traditional active learning tries to identify instances for which the acquisition of the label increases model performance under budget constraints. Less research has been devoted to the task of actively acquiring feature values, whereupon both the instance and the feature must be selected intelligently and even less to a scenario where the instances arrive in a stream with feature drift. We propose an active feature acquisition strategy for data streams with feature drift, as well as an active feature acquisition evaluation framework. We also implement a baseline that chooses features randomly and compare the random approach against eight different methods in a scenario where we can acquire at most one feature at the time per instance and where all features are considered to cost the same. Our initifal experiments on 9 different data sets, with 7 different degrees of missing features and 8 different budgets show that our developed methods outperform the random acquisition on 7 data sets and have a comparable performance on the remaining two.

Title
Active feature acquisition on data streams under feature drift
Author
Eirini Ntoutsi, Christian Beyer, Maik Büttner, Vishnu Unnikrishnan, Miro Schleicher, Eirini Ntoutsi, Myra Spiliopoulou
Publisher
Springer Link
Date
2020-07-08
Appeared in
Annals of Telecommunications, volume 75, (2020)
Citation
Beyer, C., Büttner, M., Unnikrishnan, V. et al. Active feature acquisition on data streams under feature drift. Ann. Telecommun. 75, 597–611 (2020). https://doi.org/10.1007/s12243-020-00775-2
Size or Duration
pp. 597-611
Rights
open access