Application of Gene Expression Programming in Lithology Identification and Classification

摘要

Lithology identification and classification is one of the most important issues in geological and mineral exploration. Gene expression programming (GEP) algorithm is a powerful method in data mining and pattern recognition and it has been widely used in these areas since it was proposed. In this paper, the litho-geochemical data in Ha-da-men valley, Neimeng, China, is taken for a case study to illustrate the application of GEP in lithology identification and classification. Based on a geological conceptual model, geochemical concentration values, TiO_2, FeO, MgO, K_2O and H_2O~+ are selected as input variables in the GEP model. It does not only ease to explain the geological significances, but also reduces the number of variables so as to improve the computing speed. The results show that GEP could be a powerful method to effectively identify different lithology and recognize the key variables which could show different lithological characteristics from multivariable that the geochemical concentration values in various lithological rocks. This method may also be extended to remote sensing for various ground objects recognition in mineral exploration and prospecting.