bsxfun: normalizing many vectors

Y⁽ⁱ⁾=X⁽ⁱ⁾/||X⁽ⁱ⁾||2。

e.g.

X =

     1     2     3     4     

　  5     6     7     8

X_norm = sqrt( sum(X.^2,1) );

Y = bsxfun(@rdivide, X, X_norm);

--->

X_norm =

    5.0990    6.3246    7.6158    8.9443

Y =

    0.1961    0.3162    0.3939    0.4472   

    0.9806    0.9487    0.9191    0.8944

bsxfun: normalizing many vectors

原文：http://www.cnblogs.com/deeplearning2015/p/4984155.html