## Linear Transformations and Matrices
Remember that a linear transformation $T:\mathbf{R}^{m}\to\mathbf{R}^{n}$ is a function
that satisfies the two conditions:
- $T(ax)=aT(x)$ for all $x\in\mathbf{R}^{m}$ and $a\in\mathbf{R}$.
- $T(x+y)=T(x)+T(y)$ for all $x,y\in\mathbf{R}^{m}$.
We saw earlier that a linear transformation can be represented by an $n\times m$ matrix $A$
where
$$
T(x_1,\ldots, x_m) = A\left[\begin{matrix} x_1 \\ \vdots \\ x_n\end{matrix}\right]
$$
## Linear transformations and bases
We can take a slightly more general point of view on matrices and linear transformations.
In the earlier version we used "standard coordinates" where $x_1,\ldots, x_n$ are relative to the "standard basis."
Now suppose $B=\{b_1,\ldots, b_n\}$ are a basis for $\mathbf{R}^{n}$.
Then if
$$
x=r_1 b_1 + \cdots + r_n b_n
$$
we have the coordinate vector
$$
[x]_B=\left[\begin{matrix} r_1 \\ r_2 \\ \vdots \\ r_n\end{matrix}\right].
$$
## Linear transformations in other bases
By linearity
$$
T(x) = T(r_1 b_1 + \cdots + r_n b_n)= r_1 T(b_1)+\cdots+ r_{n}T(b_n)
$$
In other words, if we make a matrix $M$ whose *columns*
are the vectors $T(b_i)$, then
$$
T(x) = M[x]_{B}=\left[\begin{matrix} r_1 \\ \vdots \\ r_n\end{matrix}\right]
$$
The matrix $M$ is called *the matrix of the linear transformation $T$ in the basis $B$* and is written
$$
M=[T]_{B}
$$
## Linear transformations and change of basis
If we write $S$ for the standard basis, the "change of basis matrix" $P_{S\leftarrow B}$ (which the book calls just $P_{B}$ has the property that
$$
P_{S\leftarrow B}[x]_{B} = [x]_{S}
$$
If $T(x) = Ax$, then in our notation above $A=[T]_{S}$ and
$x=[x]_{S}$. We can write this equation as
$$
[T(x)]_{S}=[T]_{S}[x]_{S}
$$
## Linear transformations and change of basis cont'd
So
$$
[T(x)]_{S} = A[x]_{S} = AP_{S\leftarrow B}[x]_{B}
$$
But if we want the output of $T$ to *also* be in the $B$-basis, we need one more step:
$$
[T(x)]_{B} = P_{B\leftarrow S}[T(x)]_{S}=
P_{B\leftarrow S}AP_{S\leftarrow B}[x]_{B}
$$
## Linear transformations and change of basis continued
If we simplify the notation and write $P=P_{S\leftarrow B}$
then we see that
$$
[T(x)]_{B}=[T]_{B}[x]_{B}=P^{-1}AP[x]_{B}
$$
where $A=[T]_{S}$
In other words, the matrix of $T$ in the $B$-basis
is *similar* to the matrix in the standard basis.
More generally, the collection of matrices that are similar to $A=[T]_{S}$ are the collectoin of matrix representations of $T$ in the possible bases of $\mathbf{R}^{n}$.
## Diagonal transformations
If the matrix $A$ is diagonalizable, then one can find a basis $B$
so that $[T]_{B}$ is a diagonal matrix.
## Example