loading…
A4.3.1 HL · supervised learning
Linear regression predicts a continuous value by fitting a line to the data. This activity uses gradient descent to adjust the slope and intercept. Press Play descent and watch how the loss changes.
The solid line is the model. The orange lines show its errors (residuals), the vertical gap between each point and the line. The dashed green line is the line of best fit that gradient descent is moving towards. Click to add a point.