Feature selection using stochastic diffusion search
- Submitting institution
-
University of Greenwich
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 30888
- Type
- E - Conference contribution
- DOI
-
10.1145/3071178.3079193
- Title of conference / published proceedings
- GECCO '17: Proceedings of the Genetic and Evolutionary Computation Conference
- First page
- 385
- Volume
- -
- Issue
- -
- ISSN
- -
- Open access status
- -
- Month of publication
- -
- Year of publication
- 2017
- 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)
-
-
- Citation count
- 3
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This work presents a novel application of the first swarm intelligence algorithm (Stochastic Diffusion Search, SDS) to feature selection (FS). Contrary to many swarm algorithms, SDS has a strong mathematical framework in terms of convergence to global optimum, linear complexity and minimal convergence criteria. Given the significant FS contribution to analysing high-dimensional datasets, SDS has been adapted to enable flexible determination of the desired sub-feature size. Considering the 81% outperformance of SDS-powered FS against 14 other techniques, further applications of this method to the widely required FS in high-dimensional datasets can be envisaged.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -