One-Class Support Vector Machines Based on Matrix Patterns

A Kernel-based Matrixzed One-Class Support Vector Machine

“…The decision variables in quadratic program problem (5) can be solved in an iteration way, and more elaborate description about MatOCSVM can be found in [15].…”

“…Furthermore, Wang et al improved MatLSSVM and proposed an efficient kernelized classifier named kernel-based matrixized least square support vector machine (KMatLSSVM) [14]. Yan et al [15] developed a new variant of OCSVM which directly took matrix as input, and named it Matrix-Pattern-Oriented One-class SVM (MatOCSVM). All the experiment results verify that when matrix is used as input, the matrixized classifier has a superior classification performance to its vector version.…”

Double-fold localized multiple matrixized learning machine

“…One is reducing the computational complexity and the other is improving the classification performance [7,23,43,44]. Based on the above advantages, some researchers have developed Matrix-patternoriented Ho-Kashyap classifier with regularization learning (MatMHKS) [6], New Least Squares Support Vector Classification based on matrix patterns (MatLSSVC) [36], and One-Class Support Vector Machines based on matrix patterns (OCSVM) [42].…”

Entropy-Based Support Matrix Machine

“…In order to process matrix patterns, matrix-patternoriented learning machine (MatC), i.e., matrix learning machine, has been developed. Classical learning machines include matrix-pattern-oriented Ho-Kashyap learning machine with regularization learning (MatMHKS) [2], new least squares support vector classification based on matrix patterns (MatLSSVC) [3], and oneclass support vector machines based on matrix patterns (OCSVM) [4]. Besides those matrix learning machines, Xie et.…”