Support vector machine: A tool for mapping mineral prospectivity

Abstract

In this contribution, we describe an application of support vector machine (SVM), a supervised learning algorithm, to mineral prospectivity mapping. The free R package e1071 is used to construct a SVM with sigmoid kernel function to map prospectivity for Au deposits in western Meguma Terrain of Nova Scotia (Canada). The SVM classification accuracies of ‘deposit’ are 100%, and the SVM classification accuracies of the ‘non-deposit’ are greater than 85%. The SVM classifications of mineral prospectivity have 5–9% lower total errors, 13–14% higher false-positive errors and 25–30% lower false-negative errors compared to those of the WofE prediction. The prospective target areas predicted by both SVM and WofE reflect, nonetheless, controls of Au deposit occurrence in the study area by NE–SW trending anticlines and contact zones between Goldenville and Halifax Formations. The results of the study indicate the usefulness of SVM as a tool for predictive mapping of mineral prospectivity.

Introduction

Mapping of mineral prospectivity is crucial in mineral resources exploration and mining. It involves integration of information from diverse geoscience datasets including geological data (e.g., geological map), geochemical data (e.g., stream sediment geochemical data), geophysical data (e.g., magnetic data) and remote sensing data (e.g., multispectral satellite data). These sorts of data can be visualized, processed and analyzed with the support of computer and GIS techniques. Geocomputational techniques for mapping mineral prospectivity include weights of evidence (WofE) (Bonham-Carter et al., 1989), fuzzy WofE (Cheng and Agterberg, 1999), logistic regression (Agterberg and Bonham-Carter, 1999), fuzzy logic (FL) (Ping et al., 1991), evidential belief functions (EBF) (An et al., 1992, Carranza and Hale, 2003, Carranza et al., 2005), neural networks (NN) (Singer and Kouda, 1996, Porwal et al., 2003, Porwal et al., 2004), a ‘wildcat’ method (Carranza, 2008, Carranza, 2010, Carranza and Hale, 2002) and a hybrid method (e.g., Porwal et al., 2006, Zuo et al., 2009). These techniques have been developed to quantify indices of occurrence of mineral deposit occurrence by integrating multiple evidence layers. Some geocomputational techniques can be performed using popular software packages, such as ArcWofE (a free ArcView extension) (Kemp et al., 1999), ArcSDM 9.3 (a free ArcGIS 9.3 extension) (Sawatzky et al., 2009), MI-SDM 2.50 (a MapInfo extension) (Avantra Geosystems, 2006), GeoDAS (developed based on MapObjects, which is an Environmental Research Institute Development Kit) (Cheng, 2000). Other geocomputational techniques (e.g., FL and NN) can be performed by using R and Matlab.

Geocomputational techniques for mineral prospectivity mapping can be categorized generally into two types – knowledge-driven and data-driven – according to the type of inference mechanism considered (Bonham-Carter, 1994, Pan and Harris, 2000, Carranza, 2008). Knowledge-driven techniques, such as those that apply FL and EBF, are based on expert knowledge and experience about spatial associations between mineral prospectivity criteria and mineral deposits of the type sought. On the other hand, data-driven techniques, such as WofE and NN, are based on the quantification of spatial associations between mineral prospectivity criteria and known occurrences of mineral deposits of the type sought. Additional, the mixing of knowledge-driven and data-driven methods also is used for mapping of mineral prospectivity (e.g., Porwal et al., 2006, Zuo et al., 2009). Every geocomputational technique has advantages and disadvantages, and one or the other may be more appropriate for a given geologic environment and exploration scenario (Harris et al., 2001). For example, one of the advantages of WofE is its simplicity, and straightforward interpretation of the weights (Pan and Harris, 2000), but this model ignores the effects of possible correlations amongst input predictor patterns, which generally leads to biased prospectivity maps by assuming conditional independence (Porwal et al., 2010). Comparisons between WofE and NN, NN and LR, WofE, NN and LR for mineral prospectivity mapping can be found in Singer and Kouda (1999), Harris and Pan (1999) and Harris et al. (2003), respectively.

Mapping of mineral prospectivity is a classification process, because its product (i.e., index of mineral deposit occurrence) for every location is classified as either prospective or non-prospective according to certain combinations of weighted mineral prospectivity criteria. There are two types of classification techniques. One type is known as supervised classification, which classifies mineral prospectivity of every location based on a training set of locations of known deposits and non-deposits and a set of evidential data layers. The other type is known as unsupervised classification, which classifies mineral prospectivity of every location based solely on feature statistics of individual evidential data layers.

A support vector machine (SVM) is a model of algorithms for supervised classification (Vapnik, 1995). Certain types of SVMs have been developed and applied successfully to text categorization, handwriting recognition, gene-function prediction, remote sensing classification and other studies (e.g., Joachims, 1998, Huang et al., 2002, Cristianini and Scholkopf, 2002, Guo et al., 2005, Kavzoglu and Colkesen, 2009). An SVM performs classification by constructing an n-dimensional hyperplane in feature space that optimally separates evidential data of a predictor variable into two categories. In the parlance of SVM literature, a predictor variable is called an attribute whereas a transformed attribute that is used to define the hyperplane is called a feature. The task of choosing the most suitable representation of the target variable (e.g., mineral prospectivity) is known as feature selection. A set of features that describes one case (i.e., a row of predictor values) is called a feature vector. The feature vectors near the hyperplane are the support feature vectors. The goal of SVM modeling is to find the optimal hyperplane that separates clusters of feature vectors in such a way that feature vectors representing one category of the target variable (e.g., prospective) are on one side of the plane and feature vectors representing the other category of the target variable (e.g., non-prospective) are on the other size of the plane. A good separation is achieved by the hyperplane that has the largest distance to the neighboring data points of both categories, since in general the larger the margin the better the generalization error of the classifier. In this paper, SVM is demonstrated as an alternative tool for integrating multiple evidential variables to map mineral prospectivity.