Let $A\in M(m\times n,\mathbb{K})$ be a matrix given in the row echelon form with rank $r$. Let $e_i$ be the $i$-th unit vector of $\mathbb{K}^m$. Show that $(e_1,\dots, e_r)$ is a base of $\operatorname{im}(A) \subseteq \mathbb{K}^m$

In my opinion, $A$ should look like

$$ \begin{pmatrix} a_{11} & \\ 0 & a_{22} & &*\\ \vdots & \ddots & \ddots \\ 0 & \dots & 0 & a_{rr} \end{pmatrix}$$ with $*$ I indicate arbitrarily coefficients of $A$. Because we don't change the matrix by multiplying the inverse elements to every $a_{ij}$, we obtain

$$ \begin{pmatrix} 1 & \\ 0 & 1 & &*\\ \vdots & \ddots & \ddots \\ 0 & \dots & 0 & 1 \end{pmatrix} $$ which actually are all $r$ unit vectors $(e_1,\dots,e_r)$

Since $\operatorname{im(A)}:=\operatorname{span}\{(a_{11},\dots,a_{m1}),\dots,(a_{n1},\dots,a_{mn})\}$ is the span of all column vectors of the initial matrix, we found a basis for $\operatorname{im}(A)\quad _\blacksquare$

Is this proof ok?


The matrix is an $m\times n$ matrix mapping from $\Bbb K^n\to \Bbb K^m$, it's not an $r\times r$ matrix, it just has rank $r$, so your $A$ isn't quite right (unless $r=n=m$). I'm not entirely sure what you mean by multiplying by the inverse entries of the $a_{ij}$, do you mean to multiply by the diagonal matrix $\text{diag}(a_{11},a_{22},\dots,a_{rr})$?

If you put the rank $r$ matrix $A$ in row echelon form, then with the ordered basis $(e_1,\dots,e_m)$ on $\Bbb K^m$, the image of $A$ will lie in the span of the first $r$ standard basis vectors (small example so it's easy to typeset): $$\begin{bmatrix}a&b&c&d&e\\0&f&g&h&i\\0&0&j&k&l\\0&0&0&0&0\\0&0&0&0&0\\0&0&0&0&0\end{bmatrix}.$$ (Assume the first three rows are linearly independent) This is a linear map from $\Bbb K^5\to\Bbb K^6$ and it sends $$\begin{align}e_1&\mapsto ae_1,\\e_2&\mapsto be_1+fe_2,\\e_3&\mapsto ce_1+ge_2+je_3,\\e_4&\mapsto0\end{align}$$ and so on. You can see each element of the image is written in terms of the first $r$ standard basis vectors of $\Bbb K^m$. You can lift this to the general case easily.

  • $\begingroup$ I mean that I want to multiply ever row containing some $a_{ij}$ with $\frac{1}{a_{ij}}$ $\endgroup$ – Doesbaddel Jun 14 at 10:11
  • $\begingroup$ I didn't want to express (but unfortunately I did), that we have an $r\times r$ matrix, I wanted to express, that our matrix in row echelon form has $r$ pivots that will indicate the first $r$ standard vectors. $\endgroup$ – Doesbaddel Jun 14 at 10:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.