3 ms·

[misunderstood]

• tw1010 9y ago
  You're confusing linear regression and least squares. (They're connected but not identical in that way.) Least squares gives the closest orthagonal (perpendicular) projection onto the range of the matrix. The slide is correct.