Linear regression is the simplest way for a computer to guess a number — it draws the best straight line through a bunch of dots.
Imagine it like this: Picture plotting your friends' heights against their shoe sizes on grid paper. The dots generally rise toward the right. If you draw a single straight line that passes as well as possible through all these dots, you could use it to guess a new friend's shoe size just from their height!
Examples:
So? Linear regression finds the simplest possible pattern — a straight line — that links a clue to a number to predict.