说明：   input: param: parameters of the LMGE algorithm param.mu, param.alpha, param.beta are regularization parameters. param.p: dimension of shared subspace param.k: number of nearest neighbors for Laplacian matrix X: input data Y: groundtruth labels output: model.W: matrix W Reference: Web and Personal Image Annotation by Mining Label Correlation with Relaxed Visual Graph Embedding Yi Yang, Fei Wu, Feiping Nie, Heng Tao Shen, Yueting Zhuang and Alex Hauptmann. contact: yyang@cs.cmu.edu
(Web and Personal Image Annotation by Mining Label Correlation with Relaxed Visual Graph Embedding )