A new identification method for complex lithology with support vector machine

In the research of identification method for complex lithology, cluster analysis obtains high accuracy only when the number of the samples is infinite in theory. The application of neural network is limited because of its algorithm flaw. Because Support Vector Machine (SVM) can obtain high accuracy in less samples and nonlinear multi-dimension pattern recognition problem, this method is used in order to resolve the problem of identification of complex lithology in this paper. Core data is used to establish SVM model, and this model is used to predict formation lithology in Bayantala oil filed. The correctness of lithology identification is 96%. The result shows that SVM can perfectly resolve identification problem of complex lithology.