In this lesson
- It starts with a question. Litia has a tray of mangoes already labelled ripe or unripe. The ripe ones are soft and yellow; the unripe ones are firm and green. A new mango is fairly soft and mostly yellow. With no rulebook, what is a sensible way to label it?
- The idea you will find. Label a new item by the labelled examples nearest to it. — You turned each mango into two measurements, and the new one landed beside the ripe group. That is all a nearest-neighbour classifier does: measure, find the closest labelled examples, and borrow their label. It needs trusted labels to borrow from, which is what makes this classification and not clustering.
- How you learn it. By doing: placing a point on a grid, a number line.
- Length. 6 short steps. Progress saves on this phone.
Part of Grouping & Labelling Data
Two different jobs that beginners mix up — and why the difference matters.