Application of Adaptive Neuro-Fuzzy Inference System for Grade Estimation; Case Study, Sarcheshmeh Porphyry Copper Deposit, Kerman, Iran

Knowing the grades of target elements within an explored region is a very important aspect. Such grade value properties and the element correlation is also the most important exploration parameters needed for any one who attempts to decrease the exploration cost and also the exploration risk. These could be achieved by sampling, laboratory analyses and core loggings within the boreholes. Because of the high cost and random mistakes which may happen during the sample preparation, analytical procedures and also the occurrence of other undesired events, one must be aware of getting the non-confidence results. Therefore, applying the methods which could estimate those important parameters based on the poor available information may be useful. In this paper, there is a try to apply the adaptive neuro-fuzzy inference system as a newly applied technique to solve such a problem to evaluate the "copper grade estimation" in Sarcheshmeh porphyry copper system. Based on this modeling, the input data were coordinates of samples (x, y) and the output data was the copper grade for any specific location. Training involves iterative adjustment of parameters of the adaptive neuro-fuzzy inference system using a hybrid learning procedure for mapping each training vector to its output target vector with minimum sum of squared error. The trained adaptive neuro-fuzzy inference system is used to process all feature works. A comparison of different techniques (ANN and Krining) with this new technique (ANFIS) was also carried out. The statistical parameter values of R2 in these techniques were obtained to be 0.4571 and 0.6889 respectively. After fuzzy logic and neural network combination and making an adaptive neuro-fuzzy inference system, the R2 value, changed into 0.8987. This method is expected to provide a significant improvement when the testing data come from a mixed or complex distribution