FIX : example of dense vs sparse Lasso on dense and sparse data

examples/glm/lasso_dense_vs_sparse_data.py

@@ -4,9 +4,8 @@ Lasso on dense and sparse data
 ==============================
 
 We show that glm.Lasso and glm.sparse.Lasso
-provide the same results.
-
-XXX : At the end of the day it should also lead to a speed improvement
+provide the same results and that in the case of
+sparse data glm.sparse.Lasso improves the speed.
 
 """
 
@@ -20,10 +19,10 @@ from scikits.learn.glm import Lasso as DenseLasso
 
 
 ###############################################################################
-# The two Lasso implementation on Dense data
+# The two Lasso implementations on Dense data
 print "--- Dense matrices"
 
-n_samples, n_features = 100, 10000
+n_samples, n_features = 200, 10000
 np.random.seed(0)
 y = np.random.randn(n_samples)
 X = np.random.randn(n_samples, n_features)
@@ -44,27 +43,26 @@ print "Distance between coefficients : %s" % linalg.norm(sparse_lasso.coef_
                                                         - dense_lasso.coef_)
 
 ###############################################################################
-# The two Lasso implementation on Sparse data
+# The two Lasso implementations on Sparse data
 print "--- Sparse matrices"
 
-Xs = sparse.coo_matrix(X)
-mask = Xs.data > 2 # Sparsify data matrix
-col = Xs.col[mask]
-row = Xs.row[mask]
+Xs = X.copy()
+Xs[Xs < 2.5] = 0.0
+Xs = sparse.coo_matrix(Xs)
 Xs = Xs.tocsc()
 
-print "Matrix density : %s %%" % (mask.sum() / float(X.size) * 100)
+print "Matrix density : %s %%" % (Xs.nnz / float(X.size) * 100)
 
 alpha = 0.1
 sparse_lasso = SparseLasso(alpha=alpha, fit_intercept=False)
 dense_lasso = DenseLasso(alpha=alpha, fit_intercept=False)
 
 t0 = time()
-sparse_lasso.fit(Xs, y, maxit=1000, tol=0.0)
+sparse_lasso.fit(Xs, y, maxit=1000)
 print "Sparse Lasso done in %fs" % (time() - t0)
 
 t0 = time()
-dense_lasso.fit(Xs.todense(), y, maxit=1000, tol=0.0)
+dense_lasso.fit(Xs.todense(), y, maxit=1000)
 print "Dense Lasso done in %fs" % (time() - t0)
 
 print "Distance between coefficients : %s" % linalg.norm(sparse_lasso.coef_