LavaNet - a neural network modelling toolkit for mine planning and environmental modelling

LavaNet consists of a series of scripts written in Perl that give access to a neural network simulation environment inside a general mine planning package. LavaNet enables easy development of neural network training datasets using information from any of the data and model structures available, such as block models and drillhole databases. Neural networks can be trained inside the mine planning environment and the results be used to generate new models that can be visualised in 3D. Direct comparison of developed neural network models with conventional and geostatistical techniques is now possible within the same environment. LavaNet supports Radial Basis Function networks, Multi-Layered Perceptrons and Self-Organised Maps. The flexibility of LavaNet in developing neural network based models and its integration with the mine planning environment is demonstrated using two different examples: a geological structure model and a 3D block model attribute estimation.