Support vector machines (SVMs)SVM, a set of powerful supervised machine learning algorithms was introduced by Vapnik [29, 30]. By using structural risk minimization strategy, SVMs [5] are traditionally regarded as one of the best algorithms in machine learning area. Kernel SVM works by embedding data in high dimensional or infinite feature space and an optimal separating hyperplane is constructed in the space [25]. Kernel methods for their strong capability in dealing with nonlinear data modelling, have been widely applied in the field of bioinformatics. In [15], an incremental kernel ridge regression model was proposed to predict soft-tissue deformations after craniomaxillofacial surgery.