Ore grade interpolation model based on support vector machines optimized by genetic algorithms

An approach which integrates support vector machines (SVM) and genetic algorithms (GA) was proposed to do ore grade interpolation. With the global searching characteristics, GA was used to select the optimal parameters of SVM, including the penalty parameter C, the insensitive coefficient ε and the kernel function parameter σ, so this approach can overcome the limitation of a pure SVM method in determining parameters by experience. The optimal parameters were substituted into SVM iterative training, and then an ore grade interpolation model based on SVM optimized by GA was established. Taking an underground mine in China as an example, the feasibility and validity of the ore grade interpolation model were verified by comparing the model calculation results with the actual data of mine production and the calculation results of traditional interpolation methods.