Self-implemented version of LASSO
- Accelerated by
- pre-storage
- normalization
- Resulted in only an order of magnitude slower than LASSO in sklearn
from lasso import Lasso
# general fit
lmd = 0.1
model = Lasso()
model.fit(y, X, lmd)
model.predict(X)
model.score(y, X)
# draw beta convergence path
model = Lasso()
model.path_fit(y, X)
model.draw_beta_path()