Graph embedding and extensions: a general framework for dimensionality reducation

Paper: Graph embedding and extensions: a general framework for dimensionality reducation

From：Shuicheng Yan 

N多方法来解决降维的方法：有监督和无监督的；统计的、几何理论的。。。

New 降维方法：Marginal Fisher Analysis (MFA) 

                              The intrinsic graph: characterizes the intraclass compactness and connect each data point with its neighboring points of the same class;

                              The penalty graph: connect the marginal points and characterize the interclass separability.