LearnMachine learning
Principal component analysis
Find the direction in which your data varies the most, and squeeze two dimensions into one.
Drag the round handle to turn the axis. Gold dots show where each point lands on it. Drag the pink dots to reshape the cloud. Click empty space to add a point.
How it works
Keep the direction that matters most.
Principal component analysis looks for new axes that line up with the shape of the data. The first principal component, PC1, is the direction along which the points spread out the most.
Projecting each point onto that axis keeps as much of the variation as possible with a single number. The percentage tells you how much of the total spread one axis keeps, and PC2 picks up the rest at a right angle.
With hundreds of features, a few components often keep most of the information. That makes data easier to plot, faster to learn from and less noisy, though the new axes can be harder to interpret.