Merge pull request #320 from serialhex/master

Many more ruby examples
shogun-toolbox · Aug 30, 2011 · c52988d · c52988d
2 parents faf1d9c + a42f60d
commit c52988d
Show file tree

Hide file tree

Showing 20 changed files with 624 additions and 144 deletions.
diff --git a/examples/undocumented/python_modular/classifier_domainadaptationsvm_modular.py b/examples/undocumented/python_modular/classifier_domainadaptationsvm_modular.py
@@ -14,7 +14,8 @@
 		      'CGACGTAGTCGTAGCCCCA',
 		      'CAAAAAAAAAAAAAAAATA',
 		      'CGACGGGGGGGGGGGCGTA']
-label_traindna = numpy.array(5*[-1.0] + 5*[1.0])
+label_traindna = NArray.to_na([-1.0]*5 + [1.0]*5)
+
 testdna = ['AGCACGTACGTAGCTCGAT',
 		      'AGACGTAGTCGTAGTCGTA',
 		      'CAACGGGGGGGGGGTCGTA',
@@ -25,7 +26,7 @@
 		      'CGACGTAGTCCCAGCCCCA',
 		      'CAAAAAAAAAAAACCAATA',
 		      'CGACGGCCGGGGGGGCGTA']
-label_testdna = numpy.array(5*[-1.0] + 5*[1.0])
+label_testdna = NArray.to_na([-1.0]*5 + [1.0]*5)
 
 
 traindna2 = ['AGACAGTCAGTCGATAGCT',
@@ -38,7 +39,8 @@
 		      'AGCAGTCGTAGTCGAAAAC',
 		      'ACCCCCCCCCCCCCCCCTC',
 		      'AGCAGGGGGGGGGGGAGTC']
-label_traindna2 = numpy.array(5*[-1.0] + 5*[1.0])
+label_traindna2 = NArray.to_na([-1.0]*5 + [1.0]*5)
+
 testdna2 = ['CGACAGTCAGTCGATAGCT',
 		      'CGCAGTCGTAGTCGTAGTC',
 		      'ACCAGGGGGGGGGGTAGTC',
@@ -49,39 +51,36 @@
 		      'AGCAGTCGTAAACGAAAAC',
 		      'ACCCCCCCCCCCCAACCTC',
 		      'AGCAGGAAGGGGGGGAGTC']
-label_testdna2 = numpy.array(5*[-1.0] + 5*[1.0])
+label_testdna2 = NArray.to_na([-1.0]*5 + [1.0]*5)
 
-parameter_list = [[traindna,testdna,label_traindna,label_testdna,traindna2,label_traindna2, \
-                       testdna2,label_testdna2,1,3],[traindna,testdna,label_traindna,label_testdna,traindna2,label_traindna2, \
+parameter_list = [[traindna,testdna,label_traindna,label_testdna,traindna2,label_traindna2, 
+                       testdna2,label_testdna2,1,3],[traindna,testdna,label_traindna,label_testdna,traindna2,label_traindna2, 
                        testdna2,label_testdna2,2,5]] 
 
-def classifier_domainadaptationsvm_modular(fm_train_dna=traindna,fm_test_dna=testdna, \
-                                                label_train_dna=label_traindna, \
-                                               label_test_dna=label_testdna,fm_train_dna2=traindna2,fm_test_dna2=testdna2, \
-                                               label_train_dna2=label_traindna2,label_test_dna2=label_testdna2,C=1,degree=3):
+def classifier_domainadaptationsvm_modular(fm_train_dna=traindna, fm_test_dna=testdna, label_train_dna=label_traindna, label_test_dna=label_testdna, fm_train_dna2=traindna2,fm_test_dna2=testdna2, label_train_dna2=label_traindna2, label_test_dna2=label_testdna2, c=1, degree=3)
 
 
 
-    
-	feats_train = StringCharFeatures(fm_train_dna, DNA)
-	feats_test = StringCharFeatures(fm_test_dna, DNA)
-	kernel = WeightedDegreeStringKernel(feats_train, feats_train, degree)
-	labels = Labels(label_train_dna)
-	svm = SVMLight(C, kernel, labels)
+
+	feats_train = Modshogun::StringCharFeatures.new(fm_train_dna, Modshogun::DNA)
+	feats_test = Modshogun::StringCharFeatures.new(fm_test_dna, Modshogun::DNA)
+	kernel = Modshogun::WeightedDegreeStringKernel.new(feats_train, feats_train, degree)
+	labels = Modshogun::Labels.new(label_train_dna)
+	svm = Modshogun::SVMLight.new(c, kernel, labels)
 	svm.train()
 	#svm.io.set_loglevel(MSG_DEBUG)
 
 	#####################################
 
 	#print "obtaining DA SVM from previously trained SVM"
 
-	feats_train2 = StringCharFeatures(fm_train_dna, DNA)
-	feats_test2 = StringCharFeatures(fm_test_dna, DNA)
-	kernel2 = WeightedDegreeStringKernel(feats_train, feats_train, degree)
-	labels2 = Labels(label_train_dna)
+	feats_train2 = Modshogun::StringCharFeatures.new(fm_train_dna, Modshogun::DNA)
+	feats_test2 = Modshogun::StringCharFeatures.new(fm_test_dna, Modshogun::DNA)
+	kernel2 = Modshogun::WeightedDegreeStringKernel.new(feats_train, feats_train, degree)
+	labels2 = Modshogun::Labels.new(label_train_dna)
 
 	# we regularize against the previously obtained solution
-	dasvm = DomainAdaptationSVM(C, kernel2, labels2, svm, 1.0)
+	dasvm = Modshogun::DomainAdaptationSVM.new(c, kernel2, labels2, svm, 1.0)
 	dasvm.train()
 
 	out = dasvm.apply(feats_test2).get_labels()

diff --git a/examples/undocumented/python_modular/evaluation_meansquarederror_modular.py b/examples/undocumented/python_modular/evaluation_meansquarederror_modular.py
@@ -4,20 +4,20 @@
 
 N = 100
 
-random.seed(17)
-ground_truth = random.randn(N)
-predicted = random.randn(N)
+NArray.srand(17)
+ground_truth = NArray.float(N).randomn
+predicted = NArray.float(N).randomn
 
 parameter_list = [[ground_truth,predicted]]
 
 def evaluation_meansquarederror_modular(ground_truth, predicted):
 	from shogun.Features import Labels
 	from shogun.Evaluation import MeanSquaredError
 
-	ground_truth_labels = Labels(ground_truth)
-	predicted_labels = Labels(predicted)
+	ground_truth_labels = Modshogun::Labels.new(ground_truth)
+	predicted_labels = Modshogun::Labels.new(predicted)
 
-	evaluator = MeanSquaredError()
+	evaluator = Modshogun::MeanSquaredError.new
 	mse = evaluator.evaluate(predicted_labels,ground_truth_labels)
 
 	return mse

diff --git a/examples/undocumented/python_modular/mkl_multiclass_modular.py b/examples/undocumented/python_modular/mkl_multiclass_modular.py
@@ -8,54 +8,59 @@
 		[ fm_train_real, fm_test_real, label_train_multiclass, 1.2, 1.2, 1e-5, 1, 0.001, 1.5],
 		[ fm_train_real, fm_test_real, label_train_multiclass, 5, 1.2, 1e-2, 1, 0.001, 2]]
 
-def mkl_multiclass_modular(fm_train_real, fm_test_real, label_train_multiclass,
-	width, C, epsilon, num_threads, mkl_epsilon, mkl_norm):
+def mkl_multiclass_modular(fm_train_real, fm_test_real, label_train_multiclass, width, c, epsilon, num_threads, mkl_epsilon, mkl_norm)
 
 	from shogun.Features import CombinedFeatures, RealFeatures, Labels
 	from shogun.Kernel import CombinedKernel, GaussianKernel, LinearKernel,PolyKernel
 	from shogun.Classifier import MKLMultiClass
 
-	kernel = CombinedKernel()
-	feats_train = CombinedFeatures()
-	feats_test = CombinedFeatures()
+	kernel = Modshogun::CombinedKernel.new
+	feats_train = Modshogun::CombinedFeatures.new
+	feats_test = Modshogun::CombinedFeatures.new
 
-	subkfeats_train = RealFeatures(fm_train_real)
-	subkfeats_test = RealFeatures(fm_test_real)
-	subkernel = GaussianKernel(10, width)
+	subkfeats_train = Modshogun::RealFeatures.new
+	subkfeats_train.set_feature_matrix(fm_train_real)
+	subkfeats_test = Modshogun::RealFeatures.new
+	subkfeats_test.set_feature_matrix(fm_test_real)
+	subkernel = Modshogun::GaussianKernel.new(10, width)
 	feats_train.append_feature_obj(subkfeats_train)
 	feats_test.append_feature_obj(subkfeats_test)
 	kernel.append_kernel(subkernel)
 
-	subkfeats_train = RealFeatures(fm_train_real)
-	subkfeats_test = RealFeatures(fm_test_real)
-	subkernel = LinearKernel()
+	subkfeats_train = Modshogun::RealFeatures.new
+	subkfeats_train.set_feature_matrix(fm_train_real)
+	subkfeats_test = Modshogun::RealFeatures.new
+	subkfeats_test.set_feature_matrix(fm_test_real)
+	subkernel = Modshogun::LinearKernel.new
 	feats_train.append_feature_obj(subkfeats_train)
 	feats_test.append_feature_obj(subkfeats_test)
 	kernel.append_kernel(subkernel)
 
-	subkfeats_train = RealFeatures(fm_train_real)
-	subkfeats_test = RealFeatures(fm_test_real)
-	subkernel = PolyKernel(10,2)
+	subkfeats_train = Modshogun::RealFeatures.new
+	subkfeats_train.set_feature_matrix(fm_train_real)
+	subkfeats_test = Modshogun::RealFeatures.new
+	subkfeats_test.set_feature_matrix(fm_test_real)
+	subkernel = Modshogun::PolyKernel.new(10,2)
 	feats_train.append_feature_obj(subkfeats_train)
 	feats_test.append_feature_obj(subkfeats_test)
 	kernel.append_kernel(subkernel)
 
 	kernel.init(feats_train, feats_train)
 
-	labels = Labels(label_train_multiclass)
+	labels = Modshogun::Labels.new(label_train_multiclass)
 
-	mkl = MKLMultiClass(C, kernel, labels)
+	mkl = Modshogun::MKLMultiClass.new(c, kernel, labels)
 
 	mkl.set_epsilon(epsilon);
 	mkl.parallel.set_num_threads(num_threads)
 	mkl.set_mkl_epsilon(mkl_epsilon)
 	mkl.set_mkl_norm(mkl_norm)
 
-	mkl.train()
+	mkl.train
 
 	kernel.init(feats_train, feats_test)
 
-	out =  mkl.apply().get_labels()
+	out =  mkl.apply.get_labels()
 	return out
 
 if __name__ == '__main__':

diff --git a/examples/undocumented/ruby_modular/classifier_custom_kernel_modular.rb b/examples/undocumented/ruby_modular/classifier_custom_kernel_modular.rb
@@ -0,0 +1,36 @@
+# this was trancekoded by the awesome trancekoder
+# ...and fixifikated by the awesum fixifikator
+require 'modshogun'
+require 'pp'
+
+parameter_list = [[1,7],[2,8]]
+
+def classifier_custom_kernel_modular(c=1,dim=7)
+
+	Modshogun::Math.init_random(c)
+
+	lab=sign(2*rand(dim) - 1)
+	data=rand(dim, dim)
+	symdata=data*data.T + diag(ones(dim))
+
+# *** 	kernel=CustomKernel()
+	kernel=Modshogun::CustomKernel.new
+	kernel.set_features()
+	kernel.set_full_kernel_matrix_from_full(data)
+# *** 	labels=Labels(lab)
+	labels=Modshogun::Labels.new
+	labels.set_features(lab)
+# *** 	svm=LibSVM(c, kernel, labels)
+	svm=Modshogun::LibSVM.new
+	svm.set_features(c, kernel, labels)
+	svm.train()
+	predictions =svm.apply() 
+	out=svm.apply().get_labels()
+	return svm,out
+
+end
+
+if __FILE__ == $0
+	puts 'custom_kernel'
+	classifier_custom_kernel_modular(*parameter_list[0])
+end
diff --git a/examples/undocumented/ruby_modular/clustering_hierarchical_modular.rb b/examples/undocumented/ruby_modular/clustering_hierarchical_modular.rb
@@ -0,0 +1,27 @@
+require 'modshogun'
+require 'pp'
+
+traindat = LoadMatrix.load_numbers('../data/fm_train_real.dat')
+
+parameter_list = [[traindat,3],[traindat,4]]
+
+def clustering_hierarchical_modular(fm_train=traindat,merges=3)
+
+	feats_train=Modshogun::RealFeatures.new
+	feats_train.set_feature_matrix(fm_train)
+	distance=Modshogun::EuclidianDistance.new(feats_train, feats_train)
+
+	hierarchical=Modshogun::Hierarchical.new(merges, distance)
+	hierarchical.train()
+
+	out_distance = hierarchical.get_merge_distances()
+	out_cluster = hierarchical.get_cluster_pairs()
+
+	return hierarchical,out_distance,out_cluster 
+
+end
+
+if __FILE__ == $0
+	puts 'Hierarchical'
+	pp clustering_hierarchical_modular(*parameter_list[0])
+end
diff --git a/examples/undocumented/ruby_modular/distance_braycurtis_modular.rb b/examples/undocumented/ruby_modular/distance_braycurtis_modular.rb
@@ -0,0 +1,28 @@
+require 'modshogun'
+require 'pp'
+
+traindat = LoadMatrix.load_numbers('../data/fm_train_real.dat')
+testdat = LoadMatrix.load_numbers('../data/fm_test_real.dat')
+
+parameter_list = [[traindat,testdat],[traindat,testdat]]
+
+def distance_braycurtis_modular(fm_train_real=traindat, fm_test_real=testdat)
+
+	feats_train=Modshogun::RealFeatures.new
+	feats_train.set_feature_matrix(fm_train_real)
+	feats_test=Modshogun::RealFeatures.new
+	feats_test.set_feature_matrix(fm_test_real)
+
+	distance=Modshogun::BrayCurtisDistance.new(feats_train, feats_train)
+
+	dm_train=distance.get_distance_matrix()
+	distance.init(feats_train, feats_test)
+	dm_test=distance.get_distance_matrix()
+
+	return distance,dm_train,dm_test
+end
+
+if __FILE__ == $0
+	puts 'BrayCurtisDistance'
+	pp distance_braycurtis_modular(*parameter_list[0])
+end
diff --git a/examples/undocumented/ruby_modular/distance_canberra_modular.rb b/examples/undocumented/ruby_modular/distance_canberra_modular.rb
@@ -0,0 +1,28 @@
+require 'modshogun'
+require 'pp'
+
+traindat = LoadMatrix.load_numbers('../data/fm_train_real.dat')
+testdat = LoadMatrix.load_numbers('../data/fm_test_real.dat')
+
+parameter_list = [[traindat,testdat],[traindat,testdat]]
+
+def distance_canberra_modular(fm_train_real=traindat,fm_test_real=testdat)
+
+	feats_train=Modshogun::RealFeatures.new
+	feats_train.set_feature_matrix(fm_train_real)
+	feats_test=Modshogun::RealFeatures.new
+	feats_test.set_feature_matrix(fm_test_real)
+
+	distance=Modshogun::CanberraMetric.new(feats_train, feats_train)
+
+	dm_train=distance.get_distance_matrix()
+	distance.init(feats_train, feats_test)
+	dm_test=distance.get_distance_matrix()
+	return distance,dm_train,dm_test
+
+end
+
+if __FILE__ == $0
+	puts 'CanberaMetric'
+	pp distance_canberra_modular(*parameter_list[0])
+end
diff --git a/examples/undocumented/ruby_modular/distribution_histogram_modular.rb b/examples/undocumented/ruby_modular/distribution_histogram_modular.rb
@@ -0,0 +1,40 @@
+require 'modshogun'
+require 'pp'
+
+traindna = LoadMatrix.load_dna('../data/fm_train_dna.dat')
+
+parameter_list = [[traindna, 3, 0, false], [traindna, 4, 0, false]]
+
+def distribution_histogram_modular(fm_dna=traindna, order=3, gap=0, reverse=false)
+
+	charfeat=Modshogun::StringCharFeatures.new(Modshogun::DNA)
+	charfeat.set_features(fm_dna)
+	feats=Modshogun::StringWordFeatures.new(charfeat.get_alphabet())
+	feats.obtain_from_char(charfeat, order-1, order, gap, reverse)
+
+	histo=Modshogun::Histogram.new
+	histo.set_features(feats)
+	histo.train()
+
+	histo.get_histogram()
+
+	num_examples=feats.get_num_vectors()
+	num_param=histo.get_num_model_parameters()
+	#for i in xrange(num_examples):
+	#	for j in xrange(num_param):
+	#		histo.get_log_derivative(j, i)
+
+	out_likelihood = histo.get_log_likelihood()
+	out_sample = histo.get_log_likelihood_sample()
+	return histo,out_sample,out_likelihood
+
+end
+
+###########################################################################
+# call functions
+###########################################################################
+
+if __FILE__ == $0
+	puts 'Histogram'
+	pp distribution_histogram_modular(*parameter_list[0])
+end
diff --git a/examples/undocumented/ruby_modular/features_read_svmlight_format_modular.rb b/examples/undocumented/ruby_modular/features_read_svmlight_format_modular.rb
@@ -0,0 +1,16 @@
+require 'modshogun'
+require 'pp'
+parameter_list=[['../data/train_sparsereal.light']]
+
+def features_read_svmlight_format_modular(fname)
+
+	f=Modshogun::SparseRealFeatures.new
+	lab=f.load_svmlight_file(fname)
+	f.write_svmlight_file('testwrite.light', lab)
+
+end
+
+if __FILE__ == $0
+	puts 'Reading SVMLIGHT format'
+	pp features_read_svmlight_format_modular(*parameter_list[0])
+end