Comparative Study on Three Autoencoder‐Based Deep Learning Algorithms for Geochemical Anomaly Identification
Abstract Deep autoencoder (AE) networks show a powerful ability for geochemical anomaly identification. Because of little contribution to the AE network, small probability samples (again, please check this) having comparatively high reconstructed errors can be recognized by the trained model as anom...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
American Geophysical Union (AGU)
2022-11-01
|
| Series: | Earth and Space Science |
| Subjects: | |
| Online Access: | https://doi.org/10.1029/2022EA002626 |
| _version_ | 1811305357449887744 |
|---|---|
| author | Bin Feng Lirong Chen Yongyang Xu Yu Zhang |
| author_facet | Bin Feng Lirong Chen Yongyang Xu Yu Zhang |
| author_sort | Bin Feng |
| collection | DOAJ |
| description | Abstract Deep autoencoder (AE) networks show a powerful ability for geochemical anomaly identification. Because of little contribution to the AE network, small probability samples (again, please check this) having comparatively high reconstructed errors can be recognized by the trained model as anomalous samples. However, different autoencoder networks have different abilities for anomaly identification. To test these methods for geochemical anomaly identification, we based our study on stream sediment data of the Cu‐Zn‐Ag metallogenic area in southwest Fujian province as samples. Three unsupervised deep learning models: the autoencoder (AE), multi‐convolutional autoencoder (MCAE), and fusion convolutional autoencoder (FCAE), were used to extract the combined structural, spatial distribution, and mixed features of multiple‐elements. The results showed that the anomalous area delineated by the FCAE model had the best consistency with the known copper mineral occurrences, followed by the MCAE and AE models, with area under the curve values (AUC) of 0.80, 0.78, and 0.61, respectively. FCAE and AE were insensitive to changes in convolution window size, while MCAE extracted more spatial distribution or mixed features. Overall, FCAE focused more on structural distribution or mixed features, combining the advantages of both MCAE and AE. Therefore, FCAE performed best among the three deep learning methods. This study provides a practical basis for selecting and constructing geochemical anomaly recognition models based on deep learning algorithms. |
| first_indexed | 2024-04-13T08:25:04Z |
| format | Article |
| id | doaj.art-d7ee33b9bd55486aaffa67efdf33555b |
| institution | Directory Open Access Journal |
| issn | 2333-5084 |
| language | English |
| last_indexed | 2024-04-13T08:25:04Z |
| publishDate | 2022-11-01 |
| publisher | American Geophysical Union (AGU) |
| record_format | Article |
| series | Earth and Space Science |
| spelling | doaj.art-d7ee33b9bd55486aaffa67efdf33555b2022-12-22T02:54:32ZengAmerican Geophysical Union (AGU)Earth and Space Science2333-50842022-11-01911n/an/a10.1029/2022EA002626Comparative Study on Three Autoencoder‐Based Deep Learning Algorithms for Geochemical Anomaly IdentificationBin Feng0Lirong Chen1Yongyang Xu2Yu Zhang3School of Geography and Information Engineering China University of Geosciences Wuhan ChinaChina Geological Survey Development and Research Center Beijing ChinaSchool of Geography and Information Engineering China University of Geosciences Wuhan ChinaInstitute of Geophysical and Geochemical Exploration China Academy of Geological Science Langfang ChinaAbstract Deep autoencoder (AE) networks show a powerful ability for geochemical anomaly identification. Because of little contribution to the AE network, small probability samples (again, please check this) having comparatively high reconstructed errors can be recognized by the trained model as anomalous samples. However, different autoencoder networks have different abilities for anomaly identification. To test these methods for geochemical anomaly identification, we based our study on stream sediment data of the Cu‐Zn‐Ag metallogenic area in southwest Fujian province as samples. Three unsupervised deep learning models: the autoencoder (AE), multi‐convolutional autoencoder (MCAE), and fusion convolutional autoencoder (FCAE), were used to extract the combined structural, spatial distribution, and mixed features of multiple‐elements. The results showed that the anomalous area delineated by the FCAE model had the best consistency with the known copper mineral occurrences, followed by the MCAE and AE models, with area under the curve values (AUC) of 0.80, 0.78, and 0.61, respectively. FCAE and AE were insensitive to changes in convolution window size, while MCAE extracted more spatial distribution or mixed features. Overall, FCAE focused more on structural distribution or mixed features, combining the advantages of both MCAE and AE. Therefore, FCAE performed best among the three deep learning methods. This study provides a practical basis for selecting and constructing geochemical anomaly recognition models based on deep learning algorithms.https://doi.org/10.1029/2022EA002626geochemical anomaly identificationautoencoderfeature combinationunsupervised learningspatial feature extraction |
| spellingShingle | Bin Feng Lirong Chen Yongyang Xu Yu Zhang Comparative Study on Three Autoencoder‐Based Deep Learning Algorithms for Geochemical Anomaly Identification Earth and Space Science geochemical anomaly identification autoencoder feature combination unsupervised learning spatial feature extraction |
| title | Comparative Study on Three Autoencoder‐Based Deep Learning Algorithms for Geochemical Anomaly Identification |
| title_full | Comparative Study on Three Autoencoder‐Based Deep Learning Algorithms for Geochemical Anomaly Identification |
| title_fullStr | Comparative Study on Three Autoencoder‐Based Deep Learning Algorithms for Geochemical Anomaly Identification |
| title_full_unstemmed | Comparative Study on Three Autoencoder‐Based Deep Learning Algorithms for Geochemical Anomaly Identification |
| title_short | Comparative Study on Three Autoencoder‐Based Deep Learning Algorithms for Geochemical Anomaly Identification |
| title_sort | comparative study on three autoencoder based deep learning algorithms for geochemical anomaly identification |
| topic | geochemical anomaly identification autoencoder feature combination unsupervised learning spatial feature extraction |
| url | https://doi.org/10.1029/2022EA002626 |
| work_keys_str_mv | AT binfeng comparativestudyonthreeautoencoderbaseddeeplearningalgorithmsforgeochemicalanomalyidentification AT lirongchen comparativestudyonthreeautoencoderbaseddeeplearningalgorithmsforgeochemicalanomalyidentification AT yongyangxu comparativestudyonthreeautoencoderbaseddeeplearningalgorithmsforgeochemicalanomalyidentification AT yuzhang comparativestudyonthreeautoencoderbaseddeeplearningalgorithmsforgeochemicalanomalyidentification |