Evaluating the performance of Extreme Learning Machine technique for ore grade estimation

• Autorzy: Abuntori Clara Akalanya , Al-Hassan Sulemana , Mireku-Gyimah Daniel , Ziggah Yao Yevenyo

Identyfikatory

DOI

10.46873/2300-3960.1062

• Języki publikacji: EN

Abstrakty

EN

Due to the complex geology of vein deposits and their erratic grade distributions, there is the tendency of overestimating or underestimating the ore grade. These estimated grade results determine the profitability of mining the ore deposit or otherwise. In this study, five Extreme Learning Machine (ELM) variants based on hard limit, sigmoid, triangular basis, sine and radial basis activation functions were applied to predict ore grade. The motive is that the activation function has been identified to play a key role in achieving optimum ELM performance. Therefore, assessing the extent of influence the activation functions will have on the final outputs from the ELM has some scientific value worth investigating. This study therefore applied ELMas ore grade estimator which is yet to be explored in the literature. The obtained results from the five ELM variants were analysed and compared with the state-of-the-art benchmark methods of Backpropagation Neural Network (BPNN) and Ordinary Kriging (OK). The statistical test results revealed that the ELM with sigmoid activation function (ELM-Sigmoid) was the best among all the other investigated methods (ELM-Hard limit, ELM-Triangular basis, ELM-Sine, ELM-Radial Basis, BPNN and OK). This is because the ELM-sigmoid produced the lowest MAE (0.0175), MSE (0.0005) and RMSE (0.0229) with highest R2 (91.93%) and R (95.88%) respectively. It was concluded that ELM-Sigmoid can be used by field practitioners as a reliable alternative ore grade estimation technique.

Słowa kluczowe

EN

extreme learning machine; artificial intelligence; artificial neural network; grade estimation; kriging

PL

ELM; sztuczna inteligencja; sztuczna sieć neuronowa; kriging

• Wydawca: Elsevier ; Główny Instytut Górnictwa
• Czasopismo: Journal of Sustainable Mining
• Rocznik: 2021
• Tom: Vol. 20, iss. 2
• Strony: 56--71
• Opis fizyczny: Bibliogr. 91 poz.

Twórcy

autor

Abuntori Clara Akalanya

• Department of Mining Engineering, University of Mines and Technology, Tarkwa, Ghana

autor

Al-Hassan Sulemana

• Department of Mining Engineering, University of Mines and Technology, Tarkwa, Ghana

autor

Mireku-Gyimah Daniel

• Department of Mining Engineering, University of Mines and Technology, Tarkwa, Ghana

autor

Ziggah Yao Yevenyo

• Department of Geomatic Engineering, University of Mines and Technology, Tarkwa, Ghana

Bibliografia

• [1] Sinclair AJ, Deraisme J. Geostatistical study of the eagle copper vein, northern British Columbia. Cim Bull 1974;746: 131-42.
• [2] Dominy SC, Camm GS, Phelps RFG. Narrow vein mining-A challenge to the operator. Mine Plan Equip Select 1997;97: 125-32.
• [3] Roy WD. Using geostatistics to estimate the resources of a narrow vein gold deposit61. Nova Scotia: DalTech; 2000. PhD Dissertation.
• [4] Dominy SC, Noppé MA, Annels AE. Errors and uncertainty in mineral resource and ore reserve estimation: the importance of getting it right. Explor Min Geol 2004;11(1-4):77-98. https://doi.org/10.2113/11.1-4.77.
• [5] Dominy SC, Annels AE, Camm GS, Wheeler P, Barr SP. Geology in the resource and reserve estimation of narrow vein deposits. Explor Min Geol 1999;6(4):317-33.
• [6] Journel AG. Nonparametric estimation of spatial distributions. J Int Assoc Math Geol 1983;15(3):445-68. https://doi.org/10.1007/BF01031292.
• [7] Arslan H. Spatial and temporal mapping of groundwater salinity using ordinary kriging and indicator kriging: the case of Bafra Plain, Turkey. Agric Water Manag 2012;113:57-63. https://doi.org/10.1016/j.agwat.2012.06.015.
• [8] Yates SR, Warrick AW, Myers DE. Disjunctive Kriging: 1 Overview of estimation and conditional probability. Water Resour Res 1986;22(5):623-30. https://doi.org/10.1029/WR022i005p00623.
• [9] Daya AA. Application of disjunctive kriging for estimating economic grade distribution in an iron ore deposit: a case study of the Choghart North Anomaly, Iran. J Geol Soc India 2014;83(5):567-76. https://doi.org/10.1007/s12594-014-0085-3.
• [10] Mendes MP, Ribeiro L. Nitrate probability mapping in the northern aquifer alluvial system of the river Tagus (Portugal) using Disjunctive Kriging. Sci Total Environ 2010;408(5): 1021-34. https://doi.org/10.1016/j.scitotenv.2009.10.069.
• [11] Verly G. The multigaussian approach and its applications to the estimation of local reserves. J Int Assoc Math Geol 1983; 15(2):259-86.
• [12] Afzal P, Madani N, Shahbeik S, Yasrebi AB. Multi-Gaussian kriging: a practice to enhance delineation of mineralised zones by Concentration-Volume fractal model in Dardevey iron ore deposit, SE Iran. J Geochem Explor 2015;158:10-21. https://doi.org/10.1016/j.gexplo.2015.06.011.
• [13] Journel AG. Recoverable reserves estimation-the geostatistical approach. Min Eng 1985;37(6):563-8.
• [14] Verly G, Sullivan J. Multigaussian and probability krigingsapplication to the jerritt canyon deposit. Min Eng 1985;37(6): 568-74.
• [15] Adhikary PP, Dash CJ, Bej R, Chandrasekharan H. Indicator and probability kriging methods for delineating Cu, Fe, and Mn contamination in groundwater of Najafgarh Block, Delhi, India. Environ Monit Assess 2011;176(1-4):663-76. https://doi.org/10.1007/s10661-010-1611-4.
• [16] Armstrong M, Boufassa A. Comparing the robustness of ordinary kriging and lognormal kriging: outlier resistance. Math Geol 1988;20(4):447-57. https://doi.org/10.1007/BF00892988.
• [17] Yamamoto JK. On unbiased backtransform of lognormal kriging estimates. Comput Geosci 2007;11(3):219-34. https://doi.org/10.1007/s10596-007-9046-x.
• [18] Arik A. Outlier restricted kriging: a new kriging algorithm for handling of outlier high grade data in ore reserve estimation. In: Proceedings of the 23rd international symposium on the application of compututers and operations research in the.Minerals industry. Littleton: Society of Mining Engineers; 1992. p. 181-7.
• [19] Leuangthong O, Nowak M. Dealing with high-grade data in resource estimation. J S Afr Inst Min Metall 2015;115(1): 27-36. https://doi.org/10.17159/2411-9717/2015/v115n1a3.
• [20] Wu X, Zhou Y. Reserve estimation using neural network techniques. Comput Geosci 1993;19(4):567-75. https://doi.org/10.1016/0098-3004(93)90082-G.
• [21] Li XL, Xie YL, Guo QJ, Li LH. Adaptive ore grade estimation method for the mineral deposit evaluation. Math Comput Model 2010;52(11-12):1947-56. https://doi.org/10.1016/j.mcm.2010.04.018.
• [22] Al-Alawi SM, Tawo EE. Application of artificial neural networks in mineral resource evaluation. J King Saud Univ Eng Sci 1998;10(1):127-38.
• [23] Kapageridis I, Denby B. In: Panagiotou GN, Michalakopoulos TN, editors. Ore grade estimation with modular neural network systems. A case study. Information Technology in the Mineral Industry. vol. 52. Rotterdam: AA Balkema Publ; 1998.
• [24] Kapageridis IK. Grade interpolation using radial basis function networks. Mine Plan Equip Select 2002:1-9.
• [25] Samanta B, Bandopadhyay S, Ganguli R, Dutta S. A comparative study of the performance of single neural network vs. Adaboost algorithm based combination of multiple neural networks for mineral resource estimation. J S Afr Inst Min Metall 2005;105(4):237-46.
• [26] Mahmoudabadi H, Izadi M, Menhaj MB. A hybrid method for grade estimation using genetic algorithm and neural networks. Comput Geosci 2008;13(1):91-101. https://doi.org/10.1007/s10596-008-9107-9.
• [27] Tahmasebi P, Hezarkhani A. Application of adaptive neurofuzzy inference system for grade estimation; case study, sarcheshmeh porphyry copper deposit, Kerman, Iran. Aust J Basic Appl Sci 2010;4(3):408-20.
• [28] Zhang X, Song S, Li J, Wu C. Robust LS-SVM regression for ore grade estimation in a seafloor hydrothermal sulphide deposit. Acta Oceanol Sin 2013;32(8):16-25. https://doi.org/10.1007/s13131-013-0337-x.
• [29] Badel M, Angorani S, Panahi MS. The application of median indicator kriging and neural network in modeling mixed population in an iron ore deposit. Comput Geosci 2010;37(4): 530-40. https://doi.org/10.1016/j.cageo.2010.07.009.
• [30] Singh RK, Ray D, Sarkar BC. Recurrent neural network approach to mineral deposit modelling. In: 4th int. Conference on recent advances on information Technology. 1-5; 2018.
• [31] Tahmasebi P, Hezarkhani A. Application of a modular feedforward neural network for grade estimation. Nat Resour Res 2011;20(1):25-32. https://doi.org/10.1007/s11053-011-9135-3.
• [32] Gholamnejad J, Kasmaee S, Kohsary A, Nezamolhosseini. A grade estimation of ore stockpiles by using artifi cial neural networks: case study on choghart iron mine in Iran. Int J Min Miner Eng 2012;4(1):17-25. https://doi.org/10.1504/IJMME.2012.047997.
• [33] Chatterjee S, Bandopadhyay S, Machuca D. Ore grade prediction using a genetic algorithm and clustering Based ensemble neural network model. Math Geosci 2010;42(3): 309-26. https://doi.org/10.1007/s11004-010-9264-y.
• [34] Jafrasteh B, Fathianpour N, Suàrez A. Comparison of machine learning methods for copper ore grade estimation. Comput Geosci 2018;22(5):1371-88. https://doi.org/10.1007/s10596-018-9758-0.
• [35] Chatterjee S, Bhattacherjee A A, Samanta B, Pal SK. Ore grade estimation of a limestone deposit in India using an artificial neural network. Appl GIS 2006;2(1):1-20. https://doi.org/10.2104/ag060003.
• [36] Dutta S, Bandopadhyay S, Ganguli R, Misra D. Machine learning algorithms and their application to ore reserve estimation of sparse and imprecise data. J Intell Learn Syst Appl 2010;2(2):86-96. https://doi.org/10.4236/jilsa.2010.22012.
• [37] Li XL, Li LH, Zhang BL, Guo QJ. Hybrid self-adaptive learning based particle swarm optimisation and support vector regression model for grade estimation. Neurocomputing 2013;118:179-90.
• [38] Samanta S, Bandopadhyay S, Ganguli R. Data segmentation and genetic algorithms for sparse data division in nome placer gold grade estimation using neural network and geostatistics. Explor Min Geol 2004;11(1-4):69-76.
• [39] Huang GB, Zhu QY, Siew CK. Extreme learning machine: a new learning scheme of feedforward neural networks. IEEE Int Joint Conf Neural Network 2004;2:985-90. https://doi.org/10.1109/IJCNN.2004.1380068.
• [40] Huang Y. Advances in artificial neural networks - methodological development and application. Algorithms 2009;2(3): 973-1007. https://doi.org/10.3390/algor2030973.
• [41] Adeel A, Larijani H, Javed A, Ahmadinia A. Impact of learning algorithms on random neural network based optimisation for LTE-UL systems. Netw Protoc Algorithm 2015; 7(3):157-78. https://doi.org/10.5296/npa.v7i3.8295.
• [42] Liu J, Yan K, Zhao X, Hu Y. Prediction of autogenous shrinkage of concretes by support vector machine. Int J Pavement Res Technol 2016;9(3):169-77. https://doi.org/10.1016/j.ijprt.2016.06.003.
• [43] Arthur CK, Temeng VA, Ziggah YY. Novel approach to predicting blast-induced ground vibration using Gaussian process regression. Eng Comput 2020;36(1):29-42. https://doi.org/10.1007/s00366-018-0686-3.
• [44] Huang GB, Zhu QY, Siew CK. Extreme learning machine: theory and applications. Neurocomputing 2006;70(1-3): 489-501. https://doi.org/10.1016/j.neucom.2005.12.126.
• [45] Poerio DV, Brown SD. Localized and adaptive soft sensor based on an extreme learning machine with automated selfcorrection strategies. J Chemometr 2020;34(7):1-18. https://doi.org/10.1002/cem.3088.
• [46] Dash R, Dash PK, Bisoi R. A self adaptive differential harmony search based optimised extreme learning machine for financial time series prediction. Swarm Evol Comput 2014; 19:25-42. https://doi.org/10.1016/j.swevo.2014.07.003.
• [47] Kapageridis IK, Denby B. Neural network modellig of ore grade spacial variability. In: International conference on arificial neural networks; 1998. p. 209-14.
• [48] Kapageridis IK, Denby B, Hunter G. GEMNET II e a neural ore grade estimation system. In: 29th international symposium on the application of computers and operations research in the minerals industries; 1999. 10pp.
• [49] Kapageridis IK, Denby B, Hunter G. Integration of a neural ore grade estimation tool in a 3D resource modeling package. Proc Int Jt Conf Neural Netw 1999;6:3908-12.
• [50] Kapageridis IK, Denby B, Schofield D. GEMNET II e an alternative method for grade estimation. 2000. p. 1-8.
• [51] Matías JM, Vaamonde A, Taboada J, González-Manteiga W. Comparison of kriging and neural networks with application to the exploitation of a slate mine. Math Geol 2004;36(4): 463-86.
• [52] Samanta B, Bandopadhyay S, Ganguli R, Dutta S. Sparse data division using data segmentation and kohonen network for neural network and geostatistical ore grade modeling in nome offshore placer deposit. Nat Resour Res 2004;13(3): 189-200.
• [53] Samanta B, Ganguli R, Bandopadhyay S. Comparing the predictive performance of neural networks with ordinary kriging in a bauxite deposit Comparing the predictive performance of neural networks with ordinary kriging in a bauxite deposit. Min Technol 2005;114(3):A129-39.
• [54] Badel M, Angorani S, Panahi MS. The application of median indicator kriging and neural network in modeling mixed population in an iron ore deposit. Comput Geosci 2011;37(4): 530-40. https://doi.org/10.1016/j.cageo.2010.07.009.
• [55] Guo WW. A novel application of neural networks for instant iron-ore grade estimation. Expert Syst Appl 2010;37(12): 8729-35. https://doi.org/10.1016/j.eswa.2010.06.043.
• [56] Dutta S, Misra D, Ganguli R, Samanta B, Bandopadhyay S. A hybrid ensemble model of kriging and neural network for ore grade estimation. Int J Surf Min Reclamat Environ 2006; 20(1):33-45. https://doi.org/10.1080/13895260500322236.
• [57] Tahmasebi P, Hezarkhani A. Comparison of optimised neural network with fuzzy logic for ore grade estimation. Austr J Basic Appl Sci 2010;4(5):764-72.
• [58] Tahmasebi P, Hezarkhani A. A hybrid neural networks-fuzzy logic-genetic algorithm for grade estimation. Comput Geosci 2012;42:18-27. https://doi.org/10.1016/j.cageo.2012.02.004.
• [59] Maleki S, Ramazia HR, Moradi S. Estimation of Iron concentration by using a support vector machine and an artificial neural network-the case study of the Choghart deposit southeast of Yazd, Yazd, Iran. Geopersia 2014;4(2):201-12.
• [60] Granek J. Application of machine learning algorithms to mineral prospectivity mapping. Vancouver: University of British Columbia; 2016. p. 105. https://doi.org/10.14288/1.0340340. PhD Dissertation.
• [61] Jafrasteh B, Fathianpour N. A hybrid simultaneous perturbation artificial bee colony and back-propagation algorithm for training a local linear radial basis neural network on ore grade estimation. Neurocomputing 2017;235:217-27. https://doi.org/10.1016/j.neucom.2017.01.016.
• [62] Jahangiri M, Ghavami Riabi SR, Tokhmechi B. Estimation of geochemical elements using a hybrid neural network-Gustafson- Kessel algorithm. J Min Environ 2018;9(2):499-511. https://doi.org/10.22044/jme.2017.5513.1363.
• [63] Manna B, Samanta B, Chakravarty D, Dutta D, Chowdhury A, Santra A, et al. Hyperspectral signature analysis using neural network for grade estimation of copper ore. IOP Conf Ser Earth Environ Sci 2018;169(1):1-12.
• [64] Smith AJ, Henry G, Frost-Killian S. A review of the birimian supergroup-and tarkwaian group-hosted gold deposits of Ghana. Episodes 2016;39(2):177-97. https://doi.org/10.18814/epiiugs/2016/v39i2/95775.
• [65] Milési JP, Ledru P, Ankrah P, Johan V, Marcoux E, Vinchon C. The metallogenic relationship between Birimian and Tarkwaian gold deposits in Ghana. Miner Deposita 1991; 26(3):228-38. https://doi.org/10.1007/BF00209263.
• [66] Cao J, Lin Z, Huang GB. Self-adaptive evolutionary extreme learning machine. Neural Process Lett 2012;36(3):285-305. https://doi.org/10.1007/s11063-012-9236-y.
• [67] Liang NY, Huang GB, Saratchandran P, Sundararajan NA. A fast and accurate online sequential learning algorithm for feedforward networks. IEEE Trans Neural Network 2006; 17(6):1411-23. https://doi.org/10.1109/TNN.2006.880583.
• [68] Sakthivadivel DA. Formalising the use of the activation function in neural inference. arXiv; 2021. p. 1-7. preprint arXiv: 2102.04896.
• [69] Dushkoff M, Ptucha R. Adaptive activation functions for deep networks. Electron Imag 2016;19:1-5. https://doi.org/10.2352/ISSN.2470-1173.2016.19.COIMG-149.
• [70] Wang Y, Li Y, Song Y, Rong X. The influence of the activation function in a convolution neural network model of facial expression recognition. Appl Sci 2020;10(5):1-20. https://doi.org/10.3390/app10051897.
• [71] Hayou S, Doucet A, Rousseau J. On the impact of the activation function on deep neural networks training. In: International conference on machine learning; 2019. p. 2672-80.
• [72] Karlik B, Olgac AV. Performance analysis of various activation functions in generalized MLP architectures of neural networks. Int J Artif Intell Expet Syst 2011;1(4):111-22.
• [73] Nwankpa C, Ijomah W, Gachagan A, Marshall S S. Activation Functions : comparison of trends in practice and research for deep learning1-20. arXiv; 2018. Prepr. arXiv1811.03378.
• [74] Neal RM. Connectionist learning of belief networks. Artif Intell 1992;56:71-113.
• [75] Sopena JM, Romero E, Alquezar R. Neural networks with periodic and monotonic activation functions: a comparative study in classification problems. Inst Elec Eng Conf Pub 1999; 1(470):323-8. https://doi.org/10.1049/cp:19991129.
• [76] Wong KW, Leung CS, Chang SJ. Handwritten digit recognition using multi-layer feedforward neural networks with periodic and monotonic activation functions. In: IEEE: In Object recognition supported by user interaction for service robots. 16; 2002. p. 106-9. https://doi.org/10.1109/icpr.2002.1047806 (3).
• [77] McCaughan DB. On the properties of a periodic perceptrons. In: Proceeding of international conference on neural networks. vol. 1; 1997. p. 188-93. https://doi.org/10.1007/BF01023055.
• [78] Lapedes A, Farber R. Nonlinear signal processing using neural networks: prediction and system modelling. Los Alamos National Laboratory: Los Alamos; 1987. p. 1-50.
• [79] Gashler MS, Ashmore SC. Modeling time series data with deep Fourier neural networks. Neurocomputing 2016;188: 3-11. https://doi.org/10.1016/j.neucom.2015.01.108.
• [80] Melkonian D. Similar basis function algorithm for numerical estimation of Fourier integrals. Numer Algorithm 2010;54(1): 73-100. https://doi.org/10.1007/s11075-009-9324-x.
• [81] Huang G, Mao K, Siew CK, Huang DS. Fast modular network implementation for support vector machines. IEEE Trans Neural Network 2005;16(6):1651-63. https://doi.org/10.1109/TNN.2005.857952.
• [82] Hornik K, Stinchcombe M, White H. Multilayer feedforward networks are universal approximators. Neural Network 1989; 2(5):359-66.
• [83] Arthur CK, Temeng VA, Ziggah YY. Soft computing-based technique as a predictive tool to estimate blast-induced ground vibration. J Sustain Min 2019;18(4):287-96. https://doi.org/10.1016/j.jsm.2019.10.001.
• [84] Gavin HP. The levenburg-marqurdt algorithm for nonlinear least squares curve-fitting problems. Durham: Department of Civil and Environmental Engineering, Duke University; 2020. p. 19.
• [85] Bhunia GS, Shit PK, Chattopadhyay R. Assessment of spatial variability of soil properties using geostatistical approach of lateritic soil (West Bengal, India). Annal Agra Sci 2018;16(4): 436-43. https://doi.org/10.1016/j.aasci.2018.06.003.
• [86] Mert BA, Dag A. A computer program for practical semivariogram modeling and ordinary kriging: a case study of porosity distribution in an oil field. Open Geosci 2017;9(1): 663-74. https://doi.org/10.1515/geo-2017-0050.
• [87] Sinclair AJ, Blackwell GH. Applied mineral inventory estimation. Cambridge: Cambridge University Press; 2006. p. 381.
• [88] Nezamolhosseini SA, Mojtahedzadeh SH, Gholamnejad J. The application of artificial neural networks to ore reserve estimation at choghart iron ore deposit. Anal Numeric Meth Min Eng 2017;6:73-83.
• [89] Asante-Okyere S, Shen C, Ziggah YY, Rulegeya MM, Zhu X. Investigating the predictive performance of Gaussian process regression in evaluating reservoir porosity and permeability. Energies 2018;11(12):1-13. https://doi.org/10.3390/en11123261.
• [90] Mueller AV, Hemond HF. Extended artificial neural networks: incorporation of a priori chemical knowledge enables use of ion selective electrodes for in-situ measurement of ions at environmentally relevant levels. Talanta 2013;117: 112-8. https://doi.org/10.1016/j.talanta.2013.08.045.
• [91] Moré JJ. The Levenberg-Marquardt algorithm: implementation and theory. In: Numerical analysis; 1978. p. 105-16. https://doi.org/10.1007/bfb0067700.