Table 1 The objects and motivations of the comparison

1

SRSVM versus SVM

To demonstrate the prior structural information within classes in the data is effective for classification

2

MPCSVM versus SVM

To demonstrate the discriminative information about each constrained pair in data is effective for classification

3

MPCSVM versus SRSVM

To display the discriminative information is more effective than the structures in data within classes for classification

4

HPCSVM versus MPCSVM

To demonstrate our newly-designed HPC regularization term is more reasonable than the MPC regularization term

5

HPCSVM versus LSSVM

To display our proposed HPCSVM is also better than SVM’s variant