Application of Support Vector Machine Method for Predicting Hydrocarbon in the Reservoir

This article is for analysis difficulties in hydrocarbon prediction of some deep T3X2 in central Sichuan Basin, using optimal selection of seismic attributes and prediction approach. We especially focusing on the cepstral coefficient´s extraction and optimization method, based on classification support vector machine method of structural risk minimization principle. The examples calculation and real test data of oil and gas verify the reliability of the proposed methods, which provides an effective new idea of hydrocarbon prediction for the research area.