Finding the best line

A linear model predicts with y = wx + b. w is the slope and b is the starting point.

When the slope grows, the output rises faster as the input increases. Training means finding the slope and intercept that sit closest to the data.

Candidate slopeError