Recognition and predicting lava underground based on regression support vector machine using seismic signa

Deep lava maybe contain petroleum and natural gas or CO2 gas. It is very important to find lava but the lava is in deep stratum and has complex structure and less sample data, so it is difficult to find lava stratum. There were some methods to predict where the lava is but some problem also was in them such as the precision of recognition and predicting was not high limited by the small number of sample, a regression SVM method of recognition and predicting underground lava is proposed, moreover, we propose a self-adaptive parameter adjust iterative algorithm to confirm SVM parameters, thereby enhancing the converging speed and the predicting accuracy. The prediction results of an example prove this method validity and practicability.