Research Article

Ranking Support Vector Machine with Kernel Approximation

Figure 1

Margin-maximization for linear RankSVM. Four data points have the preference and can be linearly ranked. and are the marginal distances for and .