INVESTIGATION OF STATISTICAL LEARNING THEORY PERFORMANCE ON CLASSIFICATION OF MULTIPLE THRESHOLD VALUES OF METAL CONTENT

Year 2017, Volume: 56 Issue: 4, 166 - 172, 01.12.2017

Güneş Ertunç

https://doi.org/10.30797/madencilik.391917

Abstract

The necessity of classifying the data according to the categorical variable is quite common in earth sciences. Especially in mining, classification regarding to the metal content, which is covered in the study, classification of geological zones for mineral resource estimation or classification of blocks in the mining production phase can be given as an example of classification problems. Geostatistical estimations methods such as kriging cannot be regarded as solution for classification, and in this study it is clearly shown by comparative case study example. In the study, support vector machines algorithm is coded that classifies depending upon position of the data, based on the statistical learning theory, which can classify multiple and binary classes. The parameter selection is automatically integrated into the algorithm. By using the categorical variables depending on the continuous independent variables from collected data, algorithm reveals the categories in the unknown locations by using only the distance based information. Through introduced algorithm in the study, categorical variables related to independent variables can be classified with respected to the definition of the problem.

Keywords

Classification, Machine learning, Support vector machines, Cadmium

İSTATİSTİKSEL ÖĞRENME TEORİSİ İLE METAL İÇERİĞİNİN ÇOKLU SINIR DEĞERLERİNDE SINIFLANDIRMA PERFORMANSININ İNCELENMESİ

Abstract

Verilerin kategorik değişkenliğe göre sınıflandırılması gerekliliği madencilikte oldukça sık rastlanan durumdur. Bu çalışma kapsamında ele alınan metal içeriğine göre sınıflandırma veya jeolojik zonların maden kaynak kestirimi için sınıflandırılması, madencilik üretim aşamasında
blokların sınıflandırılması örnek olarak sayılabilir. Krigleme gibi jeoistatistiksel kestirim yöntemleri, sınıflandırma için çözüm üreten bir araç değildir ve çalışmada karşılaştırmalı olarak neden kullanılmaması gerektiği açıkça ortaya konmuştur. Çalışmada, ikili sınıftan fazla, çoklu
sınıfların etkin bir şekilde sınıflandırmaya yarayan, istatistiksel öğrenme teorisine dayalı, verilerin konumuna bağlı olarak sınıflandırma yapan ve parametre seçimi otomatik halde algoritmaya entegre edilen bir destek vektör makinesi programı kodlanmıştır. Bu program sayesinde bağımsız
değişkenlere bağlı kategorik değişkenler problemin tanımına göre sınıflandırılabilmektedir. Algoritma girdisi olarak sahada toplanan verilerin devamlı bağımsız değişkenlerine göre var olan kategorik değişkenlerin, sahada bilinmeyen lokasyonlardaki kategorileri, sadece uzaklığa bağlı
konumları kullanılarak ortaya konabilmektedir.

Keywords

Sınıflandırma, Makine öğrenme, Destek vektör makineleri, Kadmiyum

References

• Armstrong, M., 1998. Basic Linear Geostatistics (Springer Berlin Heidelberg).
• Atalay F., Tercan A.E., 2017. Coal resource estimation using Gaussian copula. International Journal of Coal Geology, 175, 1-9.
• Atteia, O., Dubois J. P., Webster R., 1994. Geostatistical Analysis of Soil Contamination in the Swiss Jura. Environmental Pollution, 86: 315-27.
• Bahria, S., Essoussi N., Limam M., 2011. Hyperspectral data classification using geostatistics and support vector machines. Remote Sensing Letters, 2: 99-106.
• Bishop, C.M., 2006. Pattern Recognition and Machine Learning (Information Science and Statistics) (Springer-Verlag New York, Inc.).
• Burges, C.J.C., 1998. A Tutorial on Support Vector Machines for Pattern Recognition. Data Mining and Knowledge Discovery, 2: 121-67.
• Chilès, J.P., Delfiner P., 1999. Geostatistics: modeling spatial uncertainty (Wiley).
• Cressie, N.A.C., 1991. Statistics for Spatial Data (J. Wiley).
• Deutsch, C.V., Journel A.G., 1998. GSLIB - Geostatistical Software Library and User’s Guide (Oxford University Press).
• FOEFL (Swiss Federal Office of Environment, Forest and Landscape) 1987. Commentary on the Ordinance Relating to Pollutants in Soil (VSBo; of June 9, 1986) (FOEFL).
• Goovaerts, P., Webster R., Dubois J.P., 1997. Assessing the risk of soil contamination in the Swiss Jura using indicator geostatistics. Environmental and Ecological Statistics, 4: 49-64.
• Goovaerts, P., 1997. Geostatistics for natural resources evaluation (Oxford University Press: New York).
• Isaaks, E. H., Srivastava R.M., 1989. Applied Geostatistics (Oxford University Press).
• Kanevski, M., 1999. Spatial Predictions of Soil Contamination Using General Regression Neural Networks. Int. Journal of Systems Research and Information Systems, 8: 15.
• Kanevski, M., Canu, S., 2000b. Spatial Data Mapping with Support Vector Regression and Geostatistics. In.: IDIAP Research Report.
• Kanevski, M., Pozdnukhov A., Canu S., Maignan M., Wong P. M., Shibli S. A. R., 2002. Support Vector Machines for Classification and Mapping of Reservoir Data. in Patrick Wong, Fred Aminzadeh and Masoud Nikravesh (eds.), Soft Computing for Reservoir Characterization and Modeling (Physica-Verlag HD: Heidelberg).
• Kanevski, M., Pozdnukhov, A., Canu, S., Maignan, M., 2000a. Advanced Spatial Data Analysis and Modelling with Support Vector Machines. In.: IDIAP Research Report.
• Kecman, V., 2001. Learning and Soft Computing: Support Vector Machines, Neural Networks. and Fuzzy Logic Models (MIT Press).
• Krige, D. G., 1951. A statistical approach to some basic mine valuation problems on the Witwatersrand: J. Chem. Metal. Min. Soc. South Africa, v. 52, p. 119–139.
• Matheron, G., 1967. Kriging or polynomial interpolation procedures: Trans. Canad. Inst. Min. Metal., v. 70, p. 240–244.
• Pebesma, E. J., 2004. Multivariable geostatistics in S: the gstat package. Computers & Geosciences, 30: 683-91.
• Platt, J.C., 1999. Fast training of support vector machines using sequential minimal optimization. in Sch Bernhard, lkopf, J. C. Burges Christopher and J. Smola Alexander (eds.), Advances in kernel methods (MIT Press).
• Pozdnoukhov, A., Kanevski, K., 2006. Monitoring network optimisation for spatial data classification using support vector machines. International Journal of Environment and Pollution, 28: 465-84.
• Rifkin, R., Klautau A., 2004. In Defense of One-Vs-All Classification. J. Mach. Learn. Res., 5: 101-41.
• Tercan, A.E., Ünver B., Hindistan M.A., Ertunç G., Atalay F., Ünal S., Kıllıoğlu Y., 2013. Seam modeling and resource estimation in the coalfields of western Anatolia. International Journal of Coal Geology, 112, 1, 94–106
• Vapnik, V., 1998. Statistical learning theory (Wiley).
• Vapnik, Vladimir N., 1995. The nature of statistical learning theory (Springer-Verlag New York, Inc.).
• Webster, R., Atteia O., Dubois, J. P., 1994. Coregionalization of Trace-Metals in the Soil in the Swiss Jura. European Journal of Soil Science, 45: 205-18