Comparing machine learning methods for predicting seismic P-wave velocity on global scale
Comparing machine learning methods for predicting seismic P-wave velocity on global scale
A huge amount of ocean observation data is available. A purpose of interpreting that data in marine-geophysical applications is to find, for instance, anomalies which are the signs of reservoirs in earth layers beneath the ocean floor. In this position paper, we compare different machine learning methods to predict the overall trend of seismic P-wave velocity as a function of depth for any marine location. Our study is based on a dataset consisting of data from 333 boreholes and 38 geological and spatial predictors. Our preliminary results indicate that random forests provide best results on this dataset, but also suggest to apply data augmentation for improved results with other methods.
Anomaly detection, Machine learning, Ocean observation
CEUR Workshop Proceedings
Razeghi, Yousef
83e55269-4578-4d38-b6e5-860893a7239b
Hasselbring, Wilhelm
ee89c5c9-a900-40b1-82c1-552268cd01bd
Berndt, Christian
ae5ce3a2-189c-4d47-9ff9-2a2ea3fb5ffc
Dumke, Ines
6ecd0e54-34b8-4bbd-8724-8dcbb4b4d2b3
September 2020
Razeghi, Yousef
83e55269-4578-4d38-b6e5-860893a7239b
Hasselbring, Wilhelm
ee89c5c9-a900-40b1-82c1-552268cd01bd
Berndt, Christian
ae5ce3a2-189c-4d47-9ff9-2a2ea3fb5ffc
Dumke, Ines
6ecd0e54-34b8-4bbd-8724-8dcbb4b4d2b3
Razeghi, Yousef, Hasselbring, Wilhelm, Berndt, Christian and Dumke, Ines
(2020)
Comparing machine learning methods for predicting seismic P-wave velocity on global scale.
Corpetti, T., Lenco, D., Interdonato, R., Pham, M.-T. and Lefevre, S.
(eds.)
In Proceedings of MACLEAN 2020.
vol. 2766,
CEUR Workshop Proceedings.
6 pp
.
Record type:
Conference or Workshop Item
(Paper)
Abstract
A huge amount of ocean observation data is available. A purpose of interpreting that data in marine-geophysical applications is to find, for instance, anomalies which are the signs of reservoirs in earth layers beneath the ocean floor. In this position paper, we compare different machine learning methods to predict the overall trend of seismic P-wave velocity as a function of depth for any marine location. Our study is based on a dataset consisting of data from 333 boreholes and 38 geological and spatial predictors. Our preliminary results indicate that random forests provide best results on this dataset, but also suggest to apply data augmentation for improved results with other methods.
Text
paper4
- Version of Record
More information
Published date: September 2020
Venue - Dates:
2020 MACLEAN: MAChine Learning for EArth ObservatioN Workshop, MACLEAN 2020, , Virtual, Online, 2020-09-14 - 2020-09-18
Keywords:
Anomaly detection, Machine learning, Ocean observation
Identifiers
Local EPrints ID: 488731
URI: http://eprints.soton.ac.uk/id/eprint/488731
ISSN: 1613-0073
PURE UUID: 77d73b62-efaa-4f66-8c81-cc8478f97f25
Catalogue record
Date deposited: 04 Apr 2024 17:11
Last modified: 10 Apr 2024 02:15
Export record
Contributors
Author:
Yousef Razeghi
Author:
Wilhelm Hasselbring
Author:
Christian Berndt
Author:
Ines Dumke
Editor:
T. Corpetti
Editor:
D. Lenco
Editor:
R. Interdonato
Editor:
M.-T. Pham
Editor:
S. Lefevre
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics