IDEAS home Printed from https://ideas.repec.org/a/gvu/tcsst0/v8y2026i4id2422.html

Adaptive Group Feature Stream Selection for Dynamic Feature Stream

Author

Listed:
  • Priyadarshini I.
  • Shirish Sane

Abstract

Feature selection is a dimensionality reduction technique that identifies relevant features while discarding irrelevant and redundant ones. Traditional feature selection methods require the entire feature space to be known in advance, making them suitable for static datasets. However, in real-world applications such as healthcare monitoring, sensor networks, and image analysis, new features are generated at different time intervals. In dynamic feature streams, the entire feature space is not known beforehand. To address this challenge, this paper proposes Adaptive Group Feature Stream Selection (AGFS), an online feature selection framework for dynamic feature streams. The proposed framework extends the Alpha-Investing algorithm from processing features individually to processing features that arrive in randomly sized groups. It employs a multi-stage processing approach. First, variance filtering removes constant features. Next, Alpha-Investing-based statistical testing selects relevant feature groups. A group-size capping mechanism is introduced to prevent larger feature groups from being discarded during statistical testing. Within each selected group, features are ranked using feature-class correlation, and redundant features are removed using correlation-based redundancy removal. A data-driven threshold, computed from feature-class correlation observed so far, is introduced to help prevent unnecessary elimination of complementary features. An adaptive wealth mechanism rewards the selection of informative features. Experiments conducted on six benchmark datasets using a K-Nearest Neighbours classifier with 5-fold cross-validation, demonstrate that the proposed framework achieves higher average classification accuracy (87.36%) while maintaining a competitive number of selected features compared to existing feature selection methods. The proposed approach is also validated across multiple classifiers and its consistency across classifiers is reported.

Suggested Citation

Handle: RePEc:gvu:tcsst0:v:8:y:2026:i:4:id:2422
DOI: 10.36548/jtcsst.2026.4.002
as

Download full text from publisher

File URL: https://irojournals.com/tcsst/article/view/2422
File Function: Abstract page
Download Restriction: no

File URL: https://irojournals.com/tcsst/article/download/2422/2227
File Function: Full text
Download Restriction: no

File URL: https://libkey.io/10.36548/jtcsst.2026.4.002?utm_source=ideas
LibKey link: if access is restricted and if your library uses this service, LibKey will redirect you to where you can use your library subscription to access this item
---><---

More about this item

Keywords

;
;
;
;
;

Statistics

Access and download statistics

Corrections

All material on this site has been provided by the respective publishers and authors. You can help correct errors and omissions. When requesting a correction, please mention this item's handle: RePEc:gvu:tcsst0:v:8:y:2026:i:4:id:2422. See general information about how to correct material in RePEc.

If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about.

We have no bibliographic references for this item. You can help adding them by using this form .

If you know of missing items citing this one, you can help us creating those links by adding the relevant references in the same way as above, for each refering item. If you are a registered author of this item, you may also want to check the "citations" tab in your RePEc Author Service profile, as there may be some citations waiting for confirmation.

For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: Journal Manager (email available below). General contact details of provider: https://irojournals.com/tcsst .

Please note that corrections may take a couple of weeks to filter through the various RePEc services.

IDEAS is a RePEc service. RePEc uses bibliographic data supplied by the respective publishers.