Let $V$ a $\mathbb R-$vector space of dimension $m$ and $W$ a $\mathbb R-$vector space of dimension $n$. We denote $$\mathcal L(V,W)=\{\varphi:V\to W\mid \varphi\text{ linear}\},$$ and $$\mathcal M_{n\times m} =\{\text{matrix }m\times n\}.$$
In my course it's written $$\mathcal L(V,W)\cong \mathcal M_{n\times m}(\mathbb R).$$
I'm not sure what is really mean. For me it mean that they are isomorphic, but but since for a linear map $\varphi:V\to W$, the matrix of $\varphi$ is depending on $\varphi$, changing the basis of $V$ and/or $W$ will give an other matrix to $\varphi$, and thus the map wouldn't be injective. So does $$\mathcal L(V,W)\cong \mathcal M_{n\times m}(\mathbb R),$$ really make sense ?
I however totally agree that if $\mathcal B$ is a basis of $V$ and $\mathcal B'$ is a basis of $W$, then
\begin{align*} \mathcal L(V,W)&\longrightarrow \mathcal M_{n\times m}(\mathbb R)\\ \varphi&\longmapsto (\varphi)_{\mathcal B'\mathcal B} \end{align*} is an isomorphism.
Both $\mathcal{L}(V,W)$ and $\mathcal{M}_{n\times m}(\mathbb{R})$ are algebras over the reals. Actually, they are isomorphic algebras, and the map that you defined is such an isomorphism.