Fixed tree guided multitask logistic regression

shogun-toolbox · Jul 1, 2012 · 7765f01 · 7765f01
1 parent 82caba4
commit 7765f01
Showing 1 changed file with 2 additions and 22 deletions.
diff --git a/src/shogun/lib/slep/slep_tree_mt_lr.cpp b/src/shogun/lib/slep/slep_tree_mt_lr.cpp
@@ -30,7 +30,6 @@ slep_result_t slep_tree_mt_lr(
 	double funcp = 0.0, func = 0.0;
 
 	int n_tasks = options.n_tasks;
-	//SG_SPRINT("N tasks = %d \n", n_tasks);
 
 	int iter = 1;
 	bool done = false;
@@ -81,7 +80,7 @@ slep_result_t slep_tree_mt_lr(
 		else
 			lambda_max = findLambdaMax_mt(ATb, n_feats, n_tasks, options.ind_t, options.n_nodes);
 
-		lambda_max /= n_feats*n_tasks;
+		lambda_max /= n_vecs*n_tasks;
 
 		lambda = z*lambda_max;
 	}
@@ -144,7 +143,6 @@ slep_result_t slep_tree_mt_lr(
 	while (!done && iter <= options.max_iter) 
 	{
 		beta = (alphap-1.0)/alpha;
-		//SG_SPRINT("beta = %f \n", beta);
 
 		for (i=0; i<n_feats*n_tasks; i++)
 			s[i] = w[i] + beta*wwp[i];
@@ -154,8 +152,6 @@ slep_result_t slep_tree_mt_lr(
 		for (i=0; i<n_vecs; i++)
 			As[i] = Aw[i] + beta*(Aw[i]-Awp[i]);
 
-		//SG_SPRINT("As = %f\n",SGVector<float64_t>::dot(As,As,n_vecs));
-
 		double fun_s = 0.0;
 		for (i=0; i<n_tasks*n_feats; i++)
 			g[i] = 0.0;
@@ -181,10 +177,6 @@ slep_result_t slep_tree_mt_lr(
 		}
 		fun_s /= n_vecs;
 
-		//SG_SPRINT("fun_s = %f\n", fun_s);
-
-		//SG_SPRINT("g = %f\n", SGVector<float64_t>::dot(g,g,n_feats*n_tasks));
-
 		for (i=0; i<n_feats*n_tasks; i++)
 			wp[i] = w[i];
 
@@ -195,7 +187,7 @@ slep_result_t slep_tree_mt_lr(
 			Awp[i] = Aw[i];
 
 		int inner_iter = 1;
-		while (inner_iter < 100)
+		while (inner_iter < 1000)
 		{
 			// v = s - g / L
 			for (i=0; i<n_feats*n_tasks; i++)
@@ -210,12 +202,6 @@ slep_result_t slep_tree_mt_lr(
 			else
 				altra_mt(w.matrix, v, n_feats, n_tasks, options.ind_t, options.n_nodes, lambda/L);
 
-			//SG_SPRINT("params [%d,%d,%f,%f]\n", n_feats, n_tasks, lambda/L, options.q);
-
-			//w.display_matrix();
-			//SG_SPRINT("w = %f \n", SGVector<float64_t>::dot(w.matrix,w.matrix,n_feats*n_tasks));
-
-			// v = x - s
 			for (i=0; i<n_feats*n_tasks; i++)
 				v[i] = w[i] - s[i];
 
@@ -235,9 +221,6 @@ slep_result_t slep_tree_mt_lr(
 			}
 			fun_x /= n_vecs;
 
-			//SG_SPRINT("Aw = %f\n", SGVector<float64_t>::dot(Aw,Aw,n_vecs));
-			//c.display_vector();
-
 			double r_sum = SGVector<float64_t>::dot(v,v,n_feats*n_tasks);
 			double l_sum = fun_x - fun_s - SGVector<float64_t>::dot(v,g,n_feats*n_tasks);
 
@@ -248,8 +231,6 @@ slep_result_t slep_tree_mt_lr(
 			}
 			r_sum /= 2.0;
 
-			//SG_SPRINT("sums = [%f, %f, %f]\n", r_sum, l_sum, fun_x);
-
 			if (r_sum <= 1e-20)
 			{
 				gradient_break = true;
@@ -290,7 +271,6 @@ slep_result_t slep_tree_mt_lr(
 
 		funcp = func;
 		func = fun_x + lambda*regularizer;
-		//SG_SPRINT("Obj = %f + %f * %f = %f \n",fun_x, lambda, regularizer, func);
 
 		if (gradient_break)
 			break;