In great depth underground mining the stress accumulation in the rockmass leads to a condition known as Rockburst. Until now, no detection method has proven to be successful enough in detecting rockburst events. Because of that, a software has been developed in order to predict the probability of rockburst using as data entry the in-situ stress condition and geomechanics properties of the rockmass. This software is based on Extreme Learning Machine, a single perceptron feedforward Neuronal Network that uses random projection. The foreseen result is a detection of 90% of cases and an 85% of effectivity of rockburst quality prediction. The database has Acoustic Emission readings of different great depth mines around the globe.