Abstract
The interpretation of surface electromagnetic ellipticity measurements is treated as a pattern recognition problem vis a vis an inverse problem. A parallel distributed processing or neural network approach is used for the pattern recognition. A feed forward, backpropagation network is trained to recognize the spatial location of an anomaly given ellipticity patterns from 20 theoretical models. The network has demonstrated an ability to form abstractions and generalizations when presented with ellipticity patterns that are not part of the training set.
Original language | English (US) |
---|---|
Pages | 208-212 |
Number of pages | 5 |
DOIs | |
State | Published - 1989 |
Event | 1989 Society of Exploration Geophysicists Annual Meeting, SEG 1989 - Dallas, United States Duration: Oct 29 1989 → Nov 2 1989 |
Other
Other | 1989 Society of Exploration Geophysicists Annual Meeting, SEG 1989 |
---|---|
Country/Territory | United States |
City | Dallas |
Period | 10/29/89 → 11/2/89 |
ASJC Scopus subject areas
- Geophysics