Being part of a diagonal matrix, these only multiply particular columns of Q 1 or Q 2.33 The size of the singular value tells you exactly how much influence the corresponding rows and columns of Q 1 and Q T 2 have over the original matrix. Clearly, if a singular value is tiny, very little of the corresponding rows and columns get added into the matrix A
when it is reconstituted. It is quite common for a matrix to have some zero eigenvalues, as shown
above. These zero elements on the diagonal indicate correspond to rows or columns that give no extra information. This indicates, of course, that not all of the matrix is useful— that it is rank deficient. Or, in terms of simultaneous equations, that not all the equations are independent