Re-examination of Itakpe iron ore deposit for reserve estimation using geostatistics and artificial neural network techniques

摘要

This paper re-examines the Itakpe iron ore deposit using geostatistics and artificial neural network techniques. Set of exploration information on the deposit are used to develop ordinary kriging (OK) model that produced a minimal error. The sensitivity analysis is used to choose a multilayer perceptron (MLP) network model as the optimum network for the ANN. The OK model showed a better performance for grade estimation when compared with the MLP model. Thus, using OK, a total resource of about 12% lower than that of the conventional method, which is currently in use in Itakpe, is obtained.