A Spatially Constrained Multi-Autoencoder approach for multivariate geochemical anomaly recognition

Abstract The spatial heterogeneity of geochemical background is often ignored in geochemical anomaly recognition, leading to ineffective recognition of valuable anomalies for geochemical prospecting. In this paper, a Spatially Constrained Multi-Autoencoder (SCMA) approach is proposed to deal with such an issue in multivariate geochemical anomaly recognition, which includes two unique steps: (1) with the consideration of both chemical similarity and spatial continuity of geochemical samples, a region is divided into multiple sub-domains to discriminate the various backgrounds over space, through multivariate clustering, spatial filtering, and spatial fusion; and (2) the geochemical background of each sub-domain is learned and reconstructed by a multi-autoencoder structure, which is designed to reduce the effects of random initialization of weights in an autoencoder neural network. Finally, the anomaly score is calculated as the difference between the observed geochemical features and the reconstructed features. The performance of SCMA was demonstrated by a case study involving Cu, Mn, Pb, Zn and Fe2O3 in stream sediment samples from the Chinese National Geochemical Mapping Project, in the southwestern Fujian province of China. The results showed that the spatial domain constraining greatly improved the quality of anomaly recognition, and SCMA outperformed several existing methods in all aspects. In particular, the anomalies from SCMA were the most consistent with the known Fe deposits in the area, achieving an AUC of 0.89.

[1]  Roberto Rossi,et al.  Nonstationary Z-Score Measures , 2015, Eur. J. Oper. Res..

[2]  Q. Cheng.  Multifractality and spatial statistics , 1999 .

[3]  Shouyu Chen,et al.  Identifying geochemical anomalies associated with Au–Cu mineralization using multifractal and artificial neural network models in the Ningqiang district, Shaanxi, China , 2016 .

[4]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[5]  Renguang Zuo,et al.  A comparison study of the C–A and S–A models with singularity analysis to identify geochemical anomalies in covered areas , 2013 .

[6]  R. Zuo,et al.  Big Data Analytics of Identifying Geochemical Anomalies Supported by Machine Learning Methods , 2017, Natural Resources Research.

[7]  Friedrich Götze,et al.  When is the Student $t$-statistic asymptotically standard normal? , 1997 .

[8]  R. Zuo.  Machine Learning of Mineralization-Related Geochemical Anomalies: A Review of Potential Methods , 2017, Natural Resources Research.

[9]  Anil K. Jain,et al.  Algorithms for Clustering Data , 1988 .

[10]  Heitor Silvério Lopes,et al.  A study of deep convolutional auto-encoders for anomaly detection in videos , 2018, Pattern Recognit. Lett..

[11]  C. Vogel,et al.  Asymptotic theory of filtering for linear operator equations with discrete noisy data , 1987 .

[12]  Renguang Zuo,et al.  Recognition of geochemical anomalies using a deep autoencoder network , 2016, Comput. Geosci..

[13]  Clifford R. Stanley,et al.  Comparison of probability plots and the gap statistic in the selection of thresholds for exploration geochemistry data , 1989 .

[14]  Cheng Qiuming.  Geological Process-Based Mineral Resource Quantitative Prediction and Assessment for Makeng Type Iron Polymetallic Deposits in Fujian , 2012 .

[15]  Gerald Schaefer,et al.  Lesion border detection in dermoscopy images , 2009, Comput. Medical Imaging Graph..

[16]  Yoshua Bengio,et al.  Extracting and composing robust features with denoising autoencoders , 2008, ICML '08.

[17]  Chen Qiu-ming.  Non-Linear Mineralization Model and Information Processing Methods for Prediction of Unconventional Mineral Resources , 2003 .

[18]  Davide Cozzolino,et al.  Single-image splicing localization through autoencoder-based anomaly detection , 2016, 2016 IEEE International Workshop on Information Forensics and Security (WIFS).

[19]  Yongliang Chen.  Mineral potential mapping with a restricted Boltzmann machine , 2015 .

[20]  A. Clare,et al.  A comparison of unsupervised neural networks and k-means clustering in the analysis of multi-element stream sediment data , 2001, Geochemistry: Exploration, Environment, Analysis.

[21]  J. Winter.  Practical Assessment, Research, and Evaluation Practical Assessment, Research, and Evaluation Using the Student's t-test with extremely small sample sizes Using the Student's t-test with extremely small sample sizes , 2022 .

[22]  J. A. Hartigan,et al.  A k-means clustering algorithm , 1979 .

[23]  Q. Cheng,et al.  Conditional Independence Test for Weights-of-Evidence Modeling , 2002 .

[24]  E. Carranza,et al.  Evaluation of uncertainty in mineral prospectivity mapping due to missing evidence: A case study with skarn-type Fe deposits in Southwestern Fujian Province, China , 2015 .

[25]  K. Zou,et al.  Receiver-Operating Characteristic Analysis for Evaluating Diagnostic Tests and Predictive Models , 2007, Circulation.

[26]  Renguang Zuo,et al.  A comparative study of trend surface analysis and spectrum–area multifractal model to identify geochemical anomalies , 2015 .

[27]  Xie Xuejing,et al.  Geochemical mapping in China , 1997 .

[28]  Qiuming Cheng,et al.  Spatial characteristics of geochemical patterns related to Fe mineralization in the southwestern Fujian province (China) , 2015 .

[29]  R. Zuo.  Selection of an elemental association related to mineralization using spatial analysis , 2018 .

[30]  Tom Fawcett,et al.  An introduction to ROC analysis , 2006, Pattern Recognit. Lett..

[31]  J. C. Dunn,et al.  A Fuzzy Relative of the ISODATA Process and Its Use in Detecting Compact Well-Separated Clusters , 1973 .

[32]  David A. Yuen,et al.  Applications of Microtomography to Multiscale System Dynamics: Visualisation, Characterisation and High Performance Computation , 2013 .

[33]  E. Carranza,et al.  Application of staged factor analysis and logistic function to create a fuzzy stream sediment geochemical evidence layer for mineral prospectivity mapping , 2014 .

[34]  C. Metz.  Basic principles of ROC analysis. , 1978, Seminars in nuclear medicine.

[35]  Thomas Hofmann,et al.  Greedy Layer-Wise Training of Deep Networks , 2007 .

[36]  I. Cato,et al.  Spatial clustering of metals in the sediments of the Skagerrak/Kattegat , 1999 .

[37]  Shokri Z. Selim,et al.  K-Means-Type Algorithms: A Generalized Convergence Theorem and Characterization of Local Optimality , 1984, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[38]  Chen Qiu-ming.  Quantifying the Generalized Self-Similarity of Spatial Patterns for Mineral Resource Assessment , 2004 .

[39]  Qiuming Cheng,et al.  A spatial analysis method for geochemical anomaly separation , 1996 .

[40]  Andrew P. Bradley,et al.  The use of the area under the ROC curve in the evaluation of machine learning algorithms , 1997, Pattern Recognit..

[41]  Clemens Reimann,et al.  Factor analysis applied to regional geochemical data: problems and possibilities , 2002 .

[42]  H. E. Hawkes,et al.  Geochemistry in Mineral Exploration , 1962 .

[43]  J. Tommassen,et al.  Outer membrane PhoE protein of Escherichia coli as a carrier for foreign antigenic determinants: immunogenicity of epitopes of foot-and-mouth disease virus. , 1990, Vaccine.