Abstract: There are many factors affect the variation of groundwater level. In general the variation of groundwater level is influenced by the factors of air pressure, solid tide, rain and et al. But there is strong nonlinear relation between these factors and groundwater level. The nonlinear model of relation between the groundwater level of Chongming Middle School and the factors of air pressure, solid tide and rain is established by support vector machine method. The analogizing and forecasting data of groundwater gained by support vector machine is very good. It is far better than that by stepwise regression method. The results indicate that the support vector machine possesses extensive application prospects in processing earthquake precursory data. Some other problems relating to the practical application of the support vector machine method are also discussed in this paper.

Key words: Support vector machine, Stepwise regression, Solid tide, Nonlinear relation.