Least Angle Regression¶

Least Angle Regression¶

1. Algorithm 3.2
2. We want to change so that the prediction is closer to data , i.e., we require the change of decreases . So the change should be propto `X^T r `.
3. Why this works? It reduces the MSE.
4. LAR is similar to lasso.
5. Modified LAR Algorithm 3.2a leads to lasso result.
6. LAR(lasso) is efficient. It takes steps where lasso itself might take more than p steps.
7. LAR and lasso are almost identical if we use the geometric meaning of the algorithms. But when some coefficient crosses 0, the differences pop up.