Using machine learning for hydrocarbon prospecting in Reconcavo Basin, Brazil
Author(s)
Zhakiya, Elezhan
DownloadFull printable version (2.041Mb)
Other Contributors
Massachusetts Institute of Technology. Department of Earth, Atmospheric, and Planetary Sciences.
Advisor
Bradford Hager.
Terms of use
Metadata
Show full item recordAbstract
Machine Learning techniques are being widely used in Social Sciences to find connections amongst various variables. Machine Learning connects features across different fields that do not seem to have known mathematical relationships with each other. In natural resource prospecting, machine learning can be applied to connect geochemical, geophysical, and geological variables. However, the biggest challenge in machine learning remains obtaining the data to train the ML algorithms. Here, we have applied machine learning on data extracted from maps via image processing. While the overall accuracy of prediction remains as low as 33% at this stage, we see places where the algorithm can be improved and the accuracy increased.
Description
Thesis: S.B., Massachusetts Institute of Technology, Department of Earth, Atmospheric, and Planetary Sciences, 2016. Cataloged from PDF version of thesis. Includes bibliographical references (pages [28]).
Date issued
2016Department
Massachusetts Institute of Technology. Department of Earth, Atmospheric, and Planetary SciencesPublisher
Massachusetts Institute of Technology
Keywords
Earth, Atmospheric, and Planetary Sciences.