"Support vector machines - An emerging technique for ore grade estimation"

This paper point out the predictive performance of a relatively new learning model called the support vector machines (SVM) for spatial grade estimation of two grade attributes alumina and silica in a bauxite deposit of India. In developing the SVM model, it was trained with known input-output patterns of training data set which was derived by splitting the entire data set into training and prediction sets using Genetic algorithms (GA). The use of GA algorithms was made because it attempted to divide the data into two statistical similar subsets. The predictive performance of the SVM was also compared with the ordinary kriging technique using the various statistical indices including Mean Squared Error (MSE), Mean Absolute Error (MAE), Mean Error (ME) and the coefficient of determination (R2) on the prediction data set.