The research of reservoir parameters forecasting based on KICA and SVM

In order to improve the accuracy of prediction of reservoir parameters,the paper proposes the approach for reservoir parameters forecasting based on KICA and Support vector machine (SVM).The KICA attribute optimization technology reflects the non-linear relationship and high order statistical properties of the attributes,extract the reservoir information of mutual statistical independence which reflects the reservoir parameters of the subsurface.SVM technology based on structural risk minimization principle,which can solve problems of the nonlinear systems for the small sample,high dimensional and local minimum.KICA combined the SVM,which accurately predict the reservoir parameter distributions through the huge attribute space and less well data.Through the model and actual data, it shows that reservoir parameter prediction technology has good effect of application, and high prediction.