Closed-form expression

minimize the cost explicitly without relying on an iterative algorithm

X^TX^\theta - X^T\vec{y} = 0

number 1’s result is

d \times 1

dimension if X is

n \times d

vector (d is number of coefficient count)

Fancy but Inverse matrix is very expensive

\theta = (X^TX)^{-1}X^T\vec{y}