Another important use of the SVD is the insight it gives into how much of a matrix is
important, and how to simplify a matrix with minimal disruption. The singular values are
usually arranged in order of size, with the first, σ1 , being the largest and most significant. The corresponding columns of Q 1 and Q 2 are therefore also arranged in importance. What this means is that, while we can find the exact value of A by multiplying Q 1 ΣQ T 2 , if we removed (for example) the last columns of Q 1 and Q 2 , and the final singular value, we
would be removing the least important data. If we then multiplied these simpler matrices, we would only get an approximation to A , but one which still contained all but the most insignificant information. Exercise 6 will consider this in more detail.