Review
May 31, 2019
-
Do the invertible matrices form a subspace of the vector space by all matrices?
First remember for a matrix to be invertible, it has to be a square matrix. After lec. 8, in fact an invertible matrix must be in full rank. In full rank means a square matrix has no dependency in rows or columns. Back to the question, the answer is no. Let’s use a example to illustrate:
Both and are invertible matrices but their sum is not. We can form similar cases in matrices. Inversely, do the singular matrices form a subspace of the vector space by all matrices? The answer is no again:
-
Find the basis of null space for the following matrix without multiplying the matrix out
In fact it’s not hard to multiply these two matrices, from we see the answer is just one row adding another. But one finding is important. is an invertible matrix. So it has . Then we have the following:
That is, . The nullspace of does not affect the nullspace of at all. And is already in rref, the third and fourth column is just free variables, so the basis is
-
True or False: and share the same four subspaces:
Yes. A subspace always carry a constant in front of it like to extend the basis to be a space. If is in the subspace, so does .
-
True or False: If and share the same four subspaces, is it true that .
No. An example will be
They are both invertible matrices so the nullspace and left nullspace are just zero. And they share the column and row space. One open question you might ask is if have same four subspaces, what property do they? (I don’t know the anwer)