In view of disasters caused by rock burst becoming more and more serious in coal mine production, three
models are established for evaluation and prediction the rock burst risk based on artificial neural network. First, ten
indicators are determined which have a larger influence on rock burst. Then two back propagation network models are
trained using the original data and the processed data reduced by principal component analysis respectively. And a radial
basis function network model is also established using reduced data. Finally, the performance of three different neural
network models are analyzed and the best scheme is determined for rock burst prediction.