PCA Feature Extraction for Change Detection in Multidimensional Unlabeled Data
- Submitting institution
-
Bangor University / Prifysgol Bangor
- Unit of assessment
- 12 - Engineering
- Output identifier
- UoA12_19
- Type
- D - Journal article
- DOI
-
10.1109/TNNLS.2013.2248094
- Title of journal
- IEEE Transactions on Neural Networks and Learning Systems
- Article number
- -
- First page
- 69-80
- Volume
- 25
- Issue
- 1
- ISSN
- 2162-237X
- Open access status
- Out of scope for open access requirements
- Month of publication
- March
- Year of publication
- 2014
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
1
- Research group(s)
-
-
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Principal component analysis (PCA) is typically used to reduce data dimensionality. Contrary to the practice of retaining the most descriptive principal components, we discovered that changes in the data distribution are more pronounced in the “irrelevant” components. The retained components can be fed into any further change detector, which opens the door for creating novel, more accurate change detectors. This paper was inspirational for writing a grant proposal subsequently funded by the Leverhulme Trust (RPG_2015_188, £226K).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
