cost function - measures the accuracy of our hypothesis function hθ(x)
This function is also called the squared error cost function or mean squared error.
We want hθ(x) - y to be small.
We minimize the cost function J(θ0θ1) over θ0θ1. J(θ0θ1) is a parabolic function. Minimizing it means finding the value of θ1.