Hopefully fixed MKL linadd in SVRLight

shogun-toolbox · Jun 2, 2012 · 4b8c7d0 · 4b8c7d0
1 parent 3e1dc47
commit 4b8c7d0
Show file tree

Hide file tree

Showing 2 changed files with 71 additions and 1 deletion.
diff --git a/examples/undocumented/python_modular/mkl_linadd_regression.py b/examples/undocumented/python_modular/mkl_linadd_regression.py
@@ -0,0 +1,66 @@
+from shogun.Features import CombinedFeatures, RealFeatures, BinaryLabels, RegressionLabels
+from shogun.Kernel import CombinedKernel, PolyKernel, CustomKernel
+from shogun.Classifier import MKLClassification, MKLRegression
+from modshogun import *
+from tools.load import LoadMatrix
+lm=LoadMatrix()
+
+traindat = lm.load_numbers('../data/fm_train_real.dat')
+testdat = lm.load_numbers('../data/fm_test_real.dat')
+label_traindat = lm.load_labels('../data/label_train_twoclass.dat')
+strings1 = ["hey"]*92
+strings2 = ["yeh"]*92
+
+parameter_list = [[traindat,testdat,label_traindat],[traindat,testdat,label_traindat]]
+#    fm_train_real.shape
+#    fm_test_real.shape
+#    combined_custom()
+
+def mkl_binclass_modular (fm_train_real=traindat,fm_test_real=testdat,fm_label_twoclass = label_traindat):
+
+	sc1 = StringCharFeatures(strings1, RAWBYTE)
+	sc2 = StringCharFeatures(strings2, RAWBYTE)
+
+	sfeats1 = StringWordFeatures(RAWBYTE)
+	sfeats1.obtain_from_char(sc1,0,2,0,False)
+	skernel1 = CommWordStringKernel(10,False)
+	skernel1.init(sfeats1, sfeats1)
+	sfeats2 = StringWordFeatures(RAWBYTE)
+	sfeats2.obtain_from_char(sc2,0,2,0,False)
+	skernel2 = CommWordStringKernel(10,False)
+	skernel2.init(sfeats2, sfeats2)
+
+	ffeats = RealFeatures(traindat)
+	fkernel = LinearKernel(ffeats,ffeats)
+
+	fffeats = RealFeatures(traindat)
+	ffkernel = GaussianKernel(fffeats,fffeats,1.0)
+
+# COMBINING LINADD FEATURES/KERNELS LEAD TO FAIL
+	feats_train = CombinedFeatures()
+	feats_train.append_feature_obj(ffeats)
+	#feats_train.append_feature_obj(fffeats)
+	feats_train.append_feature_obj(sfeats2)
+
+	print feats_train.get_num_vectors()
+
+	kernel = CombinedKernel()
+	kernel.append_kernel(fkernel)
+	#kernel.append_kernel(ffkernel)
+	kernel.append_kernel(skernel2)
+	kernel.init(feats_train, feats_train)
+
+	labels = RegressionLabels(fm_label_twoclass)
+	mkl = MKLRegression()
+
+	mkl.set_mkl_norm(1) #2,3
+	mkl.set_C(1, 1)
+	mkl.set_kernel(kernel)
+	mkl.set_labels(labels)
+
+	mkl.io.enable_file_and_line()
+	mkl.io.set_loglevel(MSG_DEBUG)
+	mkl.train()
+
+if __name__=='__main__':
+	mkl_binclass_modular (*parameter_list[0])
diff --git a/src/shogun/regression/svr/SVRLight.cpp b/src/shogun/regression/svr/SVRLight.cpp
@@ -573,7 +573,11 @@ void CSVRLight::update_linear_component_mkl_linadd(
 {
 	// kernel with LP_LINADD property is assumed to have
 	// compute_by_subkernel functions
-	int32_t num         = totdoc;
+	// FIXME
+	// Divided totdoc by 2 to avoid out of bounds problem. Totdoc seems
+	// to be always multiplied by 2 in regression MKL problems 
+	// SVRLight solve (Sergey Lisitsyn)
+	int32_t num         = totdoc/2;
 	int32_t num_weights = -1;
 	int32_t num_kernels = kernel->get_num_subkernels() ;
 	const float64_t* old_beta   = kernel->get_subkernel_weights(num_weights);