2 ms·
"Next, you take the square of this mismatch and call it U_1" Each U_i is a square, so it's a sum on non-negative quantities that equals zero so that each U_i =
by joaorico 12y ago
"Next, you take the square of this mismatch and call it U_1"
Each U_i is a square, so it's a sum on non-negative quantities that equals zero so that each U_i = 0 must hold separately.
That's the cool trick :)
- pfh 12y agoYes it is, it turns a problem of solving a set of equations into a minimization problem. An algorithm to find the best fit to some data can also be used to find an exact fit if it exists, and is likely to have good numerical stability.