[results]=sg('pr_loqo', 'Var1', Var1, 'Var2', Var2)
sg('load_features', filename, feature_class, type, target[, size[, comp_features]])
sg('save_features', filename, type, target)
sg('clean_features', 'TRAIN|TEST')
[features]=sg('get_features', 'TRAIN|TEST')
sg('add_features', 'TRAIN|TEST', features[, DNABINFILE|<ALPHABET>])
sg('add_multiple_features', 'TRAIN|TEST', repetitions, features[, DNABINFILE|<ALPHABET>])
sg('add_dotfeatures', 'TRAIN|TEST', features[, DNABINFILE|<ALPHABET>])
sg('set_features', 'TRAIN|TEST', features[, DNABINFILE|<ALPHABET>][, [from_position_list|slide_window], window size, [position_list|shift], skip)
sg('set_ref_features', 'TRAIN|TEST')
sg('del_last_features', 'TRAIN|TEST')
sg('convert', 'TRAIN|TEST', from_class, from_type, to_class, to_type[, order, start, gap, reversed])
sg('reshape', 'TRAIN|TEST, num_feat, num_vec)
sg('load_labels', filename, 'TRAIN|TARGET')
sg('set_labels', 'TRAIN|TEST', labels)
[labels]=sg('get_labels', 'TRAIN|TEST')
sg('set_kernel_normalization', IDENTITY|AVGDIAG|SQRTDIAG|FIRSTELEMENT|VARIANCE|ZEROMEANCENTER, size[, kernel-specific parameters])
sg('set_kernel', type, size[, kernel-specific parameters])
sg('add_kernel', weight, kernel-specific parameters)
sg('del_last_kernel')
sg('init_kernel', 'TRAIN|TEST')
sg('clean_kernel')
sg('save_kernel', filename, 'TRAIN|TEST')
[K]]=sg('get_kernel_matrix', ['TRAIN|TEST')
sg('set_WD_position_weights', W[, 'TRAIN|TEST'])
[W]=sg('get_subkernel_weights')
sg('set_subkernel_weights', W)
sg('set_subkernel_weights_combined', W, idx)
[W]=sg('get_dotfeature_weights_combined', 'TRAIN|TEST')
sg('set_dotfeature_weights_combined', W, idx)
sg('set_last_subkernel_weights', W)
[W]=sg('get_WD_position_weights')
[W]=sg('get_last_subkernel_weights')
[W]=sg('compute_by_subkernels')
sg('init_kernel_optimization')
[W]=sg('get_kernel_optimization')
sg('delete_kernel_optimization')
sg('use_diagonal_speedup', '0|1')
sg('set_kernel_optimization_type', 'FASTBUTMEMHUNGRY|SLOWBUTMEMEFFICIENT')
sg('set_solver', 'AUTO|CPLEX|GLPK|INTERNAL')
sg('set_constraint_generator', 'LIBSVM_ONECLASS|LIBSVM_MULTICLASS|LIBSVM|SVMLIGHT|LIGHT|SVMLIGHT_ONECLASS|GPBTSVM|MPDSVM|GNPPSVM|GMNPSVM')
sg('set_prior_probs', 'pos probs, neg_probs')
sg('set_prior_probs_from_labels', 'labels')
sg('resize_kernel_cache', size)
sg('set_distance', type, data type[, distance-specific parameters])
sg('init_distance', 'TRAIN|TEST')
[D]=sg('get_distance_matrix')
[result]=sg('classify')
[result]=sg('svm_classify')
[result]=sg('classify_example', feature_vector_index)
[result]=sg('svm_classify_example', feature_vector_index)
[bias, weights]=sg('get_classifier', [index in case of MultiClassSVM])
[radi, centers|merge_distances, pairs]=sg('get_clustering')
sg('new_svm', 'LIBSVM_ONECLASS|LIBSVM_MULTICLASS|LIBSVM|SVMLIGHT|LIGHT|LIGHT_ONECLASS|SVMLIN|GPBTSVM|MPDSVM|GNPPSVM|GMNPSVM|SUBGRADIENTSVM|WDSVMOCAS|SVMOCAS|SVMSGD|SVMBMRM|SVMPERF|KERNELPERCEPTRON|PERCEPTRON|LIBLINEAR_LR|LIBLINEAR_L2|LDA|LPM|LPBOOST|SUBGRADIENTLPM|KNN')
sg('new_classifier', 'LIBSVM_ONECLASS|LIBSVM_MULTICLASS|LIBSVM|SVMLIGHT|LIGHT|LIGHT_ONECLASS|SVMLIN|GPBTSVM|MPDSVM|GNPPSVM|GMNPSVM|SUBGRADIENTSVM|WDSVMOCAS|SVMOCAS|SVMSGD|SVMBMRM|SVMPERF|KERNELPERCEPTRON|PERCEPTRON|LIBLINEAR_LR|LIBLINEAR_L2|LDA|LPM|LPBOOST|SUBGRADIENTLPM|KNN')
sg('new_regression', 'SVRLIGHT|LIBSVR|KRR')
sg('new_clustering', 'KMEANS|HIERARCHICAL')
[filename, type]=sg('load_classifier')
sg('save_classifier', filename)
[number of SVMs in MultiClassSVM]=sg('get_num_svms')
[bias, alphas]=sg('get_svm', [index in case of MultiClassSVM])
sg('set_svm', bias, alphas)
sg('set_linear_classifier', bias, w)
[objective]=sg('get_svm_objective')
[objective]=sg('compute_svm_primal_objective')
[objective]=sg('compute_svm_dual_objective')
[objective]=sg('compute_mkl_primal_objective')
[objective]=sg('compute_mkl_dual_objective')
[gap]=sg('compute_relative_mkl_duality_gap')
[gap]=sg('compute_absolute_mkl_duality_gap')
sg('do_auc_maximization', 'auc')
sg('set_perceptron_parameters', learnrate, maxiter)
sg('train_classifier', [classifier-specific parameters])
sg('train_regression')
sg('train_clustering')
sg('svm_train', [classifier-specific parameters])
sg('svm_test')
sg('svm_qpsize', size)
sg('svm_max_qpsize', size)
sg('svm_bufsize', size)
sg('c', C1[, C2])
sg('svm_epsilon', epsilon)
sg('svr_tube_epsilon', tube_epsilon)
sg('svm_nu', nu)
sg('mkl_parameters', weight_epsilon, C_MKL [, mkl_norm ])
sg('elasticnet_lambda', ent_lambda)
sg('svm_max_train_time', max_train_time)
sg('use_shrinking', enable_shrinking)
sg('use_batch_computation', enable_batch_computation)
sg('use_linadd', enable_linadd)
sg('svm_use_bias', enable_bias)
sg('mkl_use_interleaved_optimization', enable_interleaved_optimization)
sg('krr_tau', tau)
sg('add_preproc', preproc[, preproc-specific parameters])
sg('del_preproc')
sg('attach_preproc', 'TRAIN|TEST', force)
sg('clean_preproc')
sg('new_hmm', N, M)
sg('load_hmm', filename)
sg('save_hmm', filename[, save_binary])
[p, q, a, b]=sg('get_hmm')
sg('append_hmm', p, q, a, b)
sg('append_model', 'filename'[, base1, base2])
sg('set_hmm', p, q, a, b)
sg('set_hmm_as', POS|NEG|TEST)
sg('chop', chop)
sg('pseudo', pseudo)
sg('load_defs', filename, init)
[result]=sg('hmm_classify')
sg('hmm_test', output name[, ROC filename[, neglinear[, poslinear]]])
[result]=sg('one_class_linear_hmm_classify')
sg('one_class_hmm_test', output name[, ROC filename[, linear]])
[result]=sg('one_class_hmm_classify')
[result]=sg('one_class_hmm_classify_example', feature_vector_index)
[result]=sg('hmm_classify_example', feature_vector_index)
sg('output_hmm')
sg('output_hmm_defined')
[likelihood]=sg('hmm_likelihood')
sg('likelihood')
sg('save_hmm_likelihood', filename[, save_binary])
[path, likelihood]=sg('get_viterbi_path', dim)
sg('vit_def')
sg('vit')
sg('bw')
sg('bw_def')
sg('bw_trans')
sg('linear_train')
sg('save_hmm_path', filename[, save_binary])
sg('convergence_criteria', num_iterations, epsilon)
sg('normalize_hmm', [keep_dead_states])
sg('add_states', states, value)
sg('permutation_entropy', width, seqnum)
[result]=sg('relative_entropy')
[result]=sg('entropy')
sg('set_feature_matrix', features)
sg('set_feature_matrix_sparse', sp1, sp2)
sg('new_plugin_estimator', pos_pseudo, neg_pseudo)
sg('train_estimator')
sg('test_estimator')
[result]=sg('plugin_estimate_classify_example', feature_vector_index)
[result]=sg('plugin_estimate_classify')
sg('set_plugin_estimate', emission_probs, model_sizes)
[emission_probs, model_sizes]=sg('get_plugin_estimate')
sg('signals_set_model', arg1)
sg('signals_set_positions', positions)
sg('signals_set_labels', labels)
sg('signals_set_split', split)
sg('signals_set_train_mask', )
sg('signals_add_feature', feature)
sg('signals_add_kernel', kernelparam)
sg('signals_run', arg1)
sg('best_path', from, to)
[prob, path, pos]=sg('best_path_2struct', p, q, cmd_trans, seq, pos, genestr, penalties, penalty_info, nbest, content_weights, segment_sum_weights)
sg('set_plif_struct', id, name, limits, penalties, transform, min_value, max_value, use_cache, use_svm)
[id, name, limits, penalties, transform, min_value, max_value, use_cache, use_svm]=sg('get_plif_struct')
sg('precompute_subkernels')
sg('precompute_content_svms', sequence, position_list, weights)
[lin_feat]=sg('get_lin_feat')
sg('set_lin_feat', lin_feat)
sg('init_dyn_prog', num_svms)
sg('clean_up_dyn_prog')
sg('init_intron_list', start_positions, end_positions, quality)
sg('precompute_tiling_features', intensities, probe_pos, tiling_plif_ids)
sg('long_transition_settings', use_long_transitions, threshold, max_len)
sg('set_model', content_weights, transition_pointers, use_orf, mod_words)
[prob, path, pos]=sg('best_path_trans', p, q, nbest, seq_path, a_trans, segment_loss)
[p_deriv, q_deriv, cmd_deriv, penalties_deriv, my_scores, my_loss]=sg('best_path_trans_deriv', , my_path, my_pos, p, q, cmd_trans, seq, pos, genestr, penalties, state_signals, penalty_info, dict_weights, mod_words [, segment_loss, segmend_ids_mask])
[W]=sg('compute_poim_wd', max_order, distribution)
[W]=sg('get_SPEC_consensus')
[W]=sg('get_SPEC_scoring', max_order)
[W]=sg('get_WD_consensus')
[W]=sg('get_WD_scoring', max_order)
[crc32]=sg('crc', string)
sg('!', system_command)
sg('exit')
sg('quit')
sg('exec', filename)
sg('set_output', 'STDERR|STDOUT|filename')
sg('set_threshold', threshold)
sg('init_random', value_to_initialize_RNG_with)
sg('threads', num_threads)
[translation]=sg('translate_string', string, order, start)
sg('clear')
sg('tic')
sg('toc')
sg('print', msg)
sg('echo', level)
sg('loglevel', 'ALL|DEBUG|INFO|NOTICE|WARN|ERROR|CRITICAL|ALERT|EMERGENCY')
sg('syntax_highlight', 'ON|OFF')
sg('progress', 'ON|OFF')
[version]=sg('get_version')
sg('help')
sg('whos')
[results]=sg('run_python', 'Var1', Var1, 'Var2', Var2,..., python_function)
[results]=sg('run_octave', 'Var1', Var1, 'Var2', Var2,..., octave_function)
[results]=sg('run_r', 'Var1', Var1, 'Var2', Var2,..., r_function)