Normalization and Linear Regression of Data

the function calculates theta(1) and theta(2) for input data X and output data y to fit a linear function h = theta(1)*X(1) + theta(2) with minimum MSE of h - y through the given data points. Elements of theta are
determined using the gradient descent method, computed iteratively until the convergence criterion is met that is when absolute relative increment of the cost function J is less or equal to the value of tolerance tol,
where J = 1/m sum((h - y).^2);