摘要
///Firstly, using “China earthquake case” as data source, the authors built a fairly complete sample set for earthquake case study through preliminary analysis and pretreatment and introduced the combination of Rough Set and BP Neural Network to the earthquake case. Secondly, the authors selected the core abnormalities which plays a decisive role in final classification from a number of complex seismic anomaly indicators as the input by use of attribute reduction algorithm based on Rough Set, and took the discrete magnitude as the output. Furthermore, the authors built a generalized BP Neural Network model to simulate the uncertain relationship between the seismic anomaly and the earthquake. Finally, the result of simulation tests showed that the precision errors of earthquake magnitude prediction is between -0.5 and 0.5.