CHIRPS: Explaining random forest classification
- Submitting institution
-
Birmingham City University
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 11Z_OP_D2008
- Type
- D - Journal article
- DOI
-
10.1007/s10462-020-09833-6
- Title of journal
- Artificial Intelligence Review
- Article number
- -
- First page
- 5747
- Volume
- 53
- Issue
- 8
- ISSN
- 0269-2821
- Open access status
- Compliant
- Month of publication
- December
- Year of publication
- 2020
- URL
-
https://link.springer.com/article/10.1007/s10462-020-09833-6
- 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
- Explaining AI decisions individually is needed in situations where each decision is highly consequential and significant inertia impedes implementing the decision. For example, if AI decides whether to effect organisational change, each decision costs many millions. Since CHIRPS outperforms state-of-the-art, it can improve confidence in costly AI decisions above a threshold of acceptability; also, it builds a valuable catalogue of objective decisions that can explain the overall decision making process in a way that subjective decision making cannot. Together with our industrial partners, we are already finalising a proposal to implement CHIRPS in the industrial sector.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -