A graph based transductive ranking algorithm
Zhibin Pan, Xiaoyan Wei · 2012
Semi-supervised ranking is a newly developed machine learning problem. In this paper, based on the graph constructed on both labeled and unlabeled data points, we propose a novel semi-supervised ranking algorithm in the transductive setting via a semi-supervised regression model. We also derive the solution in an explicit form for this model. Experiments on two QSAR data sets demonstrate its utility and effectiveness.