Linear Regression by Gradient Descent
Click to add data points and see their linear regression. Or .
The same set of points with X and Y coordinates independently normalized to the range [-1..1]. The light lines are hypotheses tested by the gradient descent algorithm.
The least squares cost function of the normalized data. Each point corresponds to a linear equation. The brighter the point, the better it fits the data. Hover the mouse to see what equation each point corresponds to.