Automatic Composition and Optimization of Multicomponent Predictive Systems With an Extended Auto-WEKA
- Submitting institution
-
Bournemouth University
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 309112
- Type
- D - Journal article
- DOI
-
10.1109/TASE.2018.2876430
- Title of journal
- IEEE Transactions on Automation Science and Engineering
- Article number
- 0
- First page
- 946
- Volume
- 16
- Issue
- 2
- ISSN
- 1545-5955
- Open access status
- Compliant
- Month of publication
- November
- Year of publication
- 2018
- 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
- 0
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Automated machine learning (AutoML) has gained significance over the last few years. This paper presents one of the early studies into AutoML, outlining some of the difficulties with respect to selecting optimal processing pipelines from the space which grows exponentially. In particular, we have proposed an automated approach to building such well-performing pipelines, by offsetting the need for expertise with computer, contributing to democratisation of machine learning and AI technologies. We also for the first time formalised the notion of data processing pipelines using Petri nets, allowing for more rigorous analysis.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -