Building a SLR model is about finding the coefficients for the equation mentioned in Linear Regression.

Given a quality metric (examples will be using MSE), it’s possible to find the best fit for a simple linear regression using two approaches.

The mathematical approach

unaswered There’s an equation that does it, and it envolves also finding the slope of the line.

For multiple linear regression:

  • Ordinary Least Squares
    • Takes long time for large datasets

The optimization (engineering) approach

  • Gradient Descent

mathcs