shogun_example.py

try:
        from shogun.PreProc import SortWordString, SortUlongString
except ImportError:
        from shogun.Preprocessor import SortWordString, SortUlongString
from shogun.Kernel import CommWordStringKernel, CommUlongStringKernel, \
                CombinedKernel

from shogun.Features import StringWordFeatures, StringUlongFeatures, \
                StringCharFeatures, CombinedFeatures, DNA, Labels
from shogun.Classifier import MSG_INFO, MSG_ERROR
try:
        from shogun.Classifier import SVMLight
except ImportError:
        from shogun.Classifier import LibSVM


train_dna=['ACGT',
	   'TGTG',
	   'GCGC',
	   'GGGG']

train_labels=[1.0,-1.0,1.0,-1.0]

test_dna=['GGGG',
	  'GTGT',
	  'AACT',
	  'GGTG']

char_feats_train=StringCharFeatures(train_dna,DNA)
char_feats_test=StringCharFeatures(test_dna,DNA)

feats_train=StringWordFeatures(DNA)
feats_test=StringWordFeatures(DNA)

feats_train.obtain_from_char(char_feats_train,1,2,0,False)
feats_test.obtain_from_char(char_feats_test,1,2,0,False)

#preprocessing for kernel
preproc = SortWordString()
preproc.init(feats_train)
preproc.init(feats_test)
feats_train.add_preproc(preproc)
feats_train.apply_preproc()
feats_test.add_preproc(preproc)
feats_test.apply_preproc()

#KERNEL
spectrum_kernel=CommWordStringKernel(feats_train,feats_train)


#svm initialization
svm=SVMLight(options.svmC, spectrum_kernel, Labels(numpy.array(labels, dtype=numpy.double)))

#use get_labels to obtain scores for test sequences
spectrum_kernel.init(feats_train,feats_test)
svm.apply().get_labels()