Naive implementation of Least Angle Regression
Latest commit a38394a Jul 19, 2013 Leif Johnson Change default learning parameters, protect from floating point excep…


A literal implementation of the LARS algorithm described by Efron, Hastie, Johnstone, and Tibshirani (2004). This implementation is much less complex than the one in scikits.learn, which I think might be useful for pedagogical purposes. For doing "real" regression problems, though, I'd recommend using the scikits.learn implementation.


Install with pip

pip install lmj.lars

Or directly from the source

git clone
cd py-lars
python install