An extensive evaluation of seven machine learning methods for rainfall prediction in weather derivatives
- Submitting institution
-
The University of Essex
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 1436
- Type
- D - Journal article
- DOI
-
10.1016/j.eswa.2017.05.029
- Title of journal
- Expert Systems with Applications
- Article number
- -
- First page
- 169
- Volume
- 85
- Issue
- -
- ISSN
- 0957-4174
- Open access status
- Compliant
- Month of publication
- May
- 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
-
3
- Research group(s)
-
A - Artificial Intelligence (AI)
- Citation count
- 45
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- In this well-cited paper in ESWA, one of the top-ranked journals, we present a novel methodology for predicting rainfall occurrence for weather derivatives. Most significantly this work is the first extensive use of machine learning algorithms on weather derivatives. We evaluate our methodology by applying 7 different ML algorithms on 42 cities, using 20 years' worth of data. All parameters are rigorously tuned via the iRace package. Results demonstrate that our approach reduces predictive error on average by 70%. This result is further supported through the well-known Friedman statistical test. The paper was foundational in a series on weather derivatives.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -