LA-2-vars-3

Read the whole caption before you watch.

The method of least squares.

A a matrix, Ax=b not solvable in x-space, i.e. b is not in R(A).

Minimize |y-b| over y in R(A).

We can do that with a projection P on R(A).

P can be computed from the consistency conditions and the GS method.

Instead we use the orthogonality condition.

We find a linear equation for the minimizer.