SVM Model Based on Signal Transformation and its Applications in Oil Water-Flooded Identification

摘要

this paper, a method of signal transformation for feature extraction is proposed. It can transform log-signal space into the vector space, which the experiment system requires, and then use SVM (Support Vector Machine) automatically to identify the water-flooded status of oil-saturated stratum. The results of experiment indicate that this algorithm has good identification ability and strong generalization ability in condition that the number of training swatch is limited.