Application of relevance vector machine to predict rockburst

摘要

Rockburst is a complex, nonlinear dynamic mechanics phenomena and its mechanism is very complex. Using monitored data for analysis and prediction of rockburst is an important research method. However, it is difficult to present the complex, nonlinear relationship between rock burst and its influence factor using conventional mathematics, mechanics methods. Rockburst forecast on the basis of time series anal-ysis is a key research direction. For the time series analysis problem of rockburst, the rockburst is regarded as a time series process, then build the nonlinear relationship between time series using relevance vector machine. RVM based rockburst prediction method is proposed. The proposed method is applied to forecast the rockburst. The results show this method is scientific, feasible and precise.