Detection of the gas-bearing zone in a carbonate reservoir using multi-class relevance vector machines (RVM): comparison of its performance with SVM and PNN

Automatic classification of seismic reflection data plays an important role in reservoir characterization and petroleum geosciences. Recently, the relevance vector machines (RVM) have attracted substantial interest in data classification literature. The RVM is a Bayesian variant of support vector machine (SVM) that can surmount the limitations of other conventional methods. The main purpose of this study is to show the effectiveness of RVM as a linear classifier to classify the variation in seismic data. In this study, we explored the potentials of RVM in identifying anomalous seismic reflection signals. First, the efficiency of RVM technique is compared with other conventional methods (SVM and probabilistic neural network (PNN)), and then a multi-class RVM algorithm based on the Gaussian approximation model is used to successfully classify synthetic and field seismic data. The result of this study shows that RVM is successful in the determination of gas-bearing zone in one of the Iranian southern oil fields.