Comparison of the Data-Driven Random Forests Model and a Knowledge-Driven Method for Mineral Prospectivity Mapping: A Case Study for Gold Deposits Around the Huritz Group and Nueltin Suite, Nunavut, Canada

Abstract

This paper outlines the process taken to create two separate gold prospectivity maps. The first was created using a combination of several knowledge-driven (KD) techniques. The second was created using a relatively new classification method called random forests (RF). The purpose of this study was to examine the results of the RF technique and to compare the results to that of the KD model. The datasets used for the creation of evidence maps for the gold prospectivity mapping include a comprehensive lake sediment geochemical dataset, interpreted geological structures (form lines), mapped and interpreted faults, lithology, topographic features (lakes), and known Au occurrences. The RF method performed well in that the gold prospectivity map created was a better predictor of the known Au occurrences than the KD gold prospectivity map. This was further validated by a fivefold repetition using a subset of the input training areas. Several advantages to the use of RF include (1) the ability to take both continuous and/or categorical data as variable inputs, (2) an internal, unbiased estimation of the mapping error (out-of-bag error) removing the need for a cross-validation of the final outputs to determine accuracy, and (3) the estimation of importance of each input variable. Efficiency of prediction curves illustrates that the RF method performs better than the KD method. The success rate is significantly higher for the RF method than for the KD method.

Testing a New Ensemble Model Based on SVM and Random Forest in Forest Fire Susceptibility Assessment and Its Mapping in Serbia’s Tara National Park

“…In this research, we carefully combine mentioned machine learning models to get an ensemble model using Bayesian averaging [38,39] with efficient feature selection to address these issues and mitigate their effects on the defect classification performance. Multiple predictions are made for each data point in Bayesian averaging.…”

Prospectivity analysis of orogenic gold deposits in Saqez-Sardasht Goldfield, Zagros Orogen, Iran

“…In MPM, mathematical functions, have been widely used to assign weights to discretized spatial evidence values as fuzzified evidential maps in the [0,1] range or to rank target areas as fuzzy prospectivity models (e.g., Bonham-Carter, 1994;Carranza and Hale, 2002;Luo and Dimitrakopoulos, 2003;Porwal et al, 2003c;Carranza, 2008Carranza, , 2009Carranza, , 2017Lisitsin et al, 2013;Mutele et al, 2017;Nykänen et al, 2017). The weights assigned to classes of discretized evidential values may be based on (a) expert judgment directly, (b) locations of known mineral occurrences (KMOs), (c) a combination of (a) and (b), or (d) subjectively-defined functions, so indirectly-assigned by analyst (e.g., Luo, 1990;Bonham-Carter, 1994;Cheng and Agterberg, 1999;Luo and Dimitrakopoulos, 2003;Porwal et al, 2003a Porwal et al, ,b,c, 2004Porwal et al, , 2006Carranza et al, 2005;Carranza, 2008Carranza, , 2014Porwal and Kreuzer, 2010;Mejía-Herrera et al, 2014;Carranza and Laborte, 2016;McKay and Harris, 2016). All these methods impart bias due to discretization of continuous spatial values, use of subjective expert judgments, and sparse or incomplete data on locations of KMOs in knowledge-and data-driven MPM (Coolbaugh et al, 2007;Lusty et al, 2012;Ford et al, 2016).…”

Tectonic environments of South American porphyry copper magmatism through time revealed by spatiotemporal data mining

“…Since random forests use bagging, it is quite easy to determine which features are important for prediction (Figure ); the exact method for this is discussed in Breiman []. This method can work well with unbalanced class data such as ours [ Oshiro et al ., ] and has been used for both mineral exploration [ Cracknell et al ., ; Carranza and Laborte , ; McKay and Harris , ] and lithological classification [ Cracknell and Reading ; Waske et al ., ].…”