Improved data visualisation through multiple dissimilarity modelling
- Submitting institution
-
Birmingham City University
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 11Z_OP_D2033
- Type
- D - Journal article
- DOI
-
10.1016/j.ins.2016.07.073
- Title of journal
- Information Sciences
- Article number
- -
- First page
- 288-302
- Volume
- 370-371
- Issue
- -
- ISSN
- 0020-0255
- Open access status
- Technical exception
- Month of publication
- -
- Year of publication
- 2016
- 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
-
-
- Research group(s)
-
-
- Citation count
- 5
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper is a novel approach to traditional dimension reduction which has been applied in a range of application domains. The removal of restrictive assumptions over data distribution and behaviour is an important step forward, allowing these techniques to be used in fields where no prior domain knowledge exists. The experiments presented utilise standard machine learning datasets to evidence rigour and clear understanding of the benefits of the techniques. This paper is built upon in a subsequent publication utilising a nonlinear interpolation of dissimilarity dictionaries allowing for chart construction on non-Riemannian manifolds currently approximated with topology-naïve deep learning models.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -