Yu Lin

with rectangular matrices, the key is almost always to consider A'A and AA'

concepts of independent/uncorrelatedness, different order of moments of join PDF

ellipsoid representation from Eigen-decomposition/SVD,used in data analysis (multivariate statistics);transormations(scale,rotate,translate)

interesting connection: SVD - pseudo inverse - least square - condition matrix - sigularity in robot workspace

ranking *****linear algebra,matrix theory(base of control,robotics,FEA.. ****probability,statistics,stochastic process ***numerical analysis

if you want to do grad studies in sci & eng, learn math as deep/broad as you could. math is indeed the foundation of all applied theory.