In this communication, experimental dissociation data for carbon dioxide, methane, and ethane hydrates in the presence of pure water and different types of porous media (mesoporous silica gel, porous glass, and silica gel) were assessed, in order to detect any suspect or doubtful data. The Leverage statistical model algorithm was used in which determination of the statistical hat matrix, sketching of the Williams plot, and calculation of the residuals of selected correlation results are included. Gas hydrate dissociation conditions in porous media were predicted using of the least squares support vector machine (LSSVM) algorithm. This study provides an overview on the quality of existing experimental data (outlier detection), as well as the applicability domain of the model employed. (C) 2013 Elsevier B.V. All rights reserved.