deleting and testing

applications/classification/random_fourier_classification.cpp

@@ -1,10 +1,7 @@
 /*
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 3 of the License, or
- * (at your option) any later version.
+ * This software is distributed under BSD 3-clause license (see LICENSE file).
  *
- * Written (W) 2013 Evangelos Anagnostopoulos
+ * Authors: Björn Esser, Evangelos Anagnostopoulos
  */
 #include <shogun/base/init.h>
 #include <shogun/features/RandomFourierDotFeatures.h>

benchmarks/hasheddoc_benchmarks.cpp

@@ -1,11 +1,7 @@
 /*
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 3 of the License, or
- * (at your option) any later version.
+ * This software is distributed under BSD 3-clause license (see LICENSE file).
  *
- * Written (W) 2013 Evangelos Anagnostopoulos
- * Copyright (C) 2013 Evangelos Anagnostopoulos
+ * Authors: Evangelos Anagnostopoulos
  */
 
 #include <shogun/base/init.h>

benchmarks/sparse_test.cpp

@@ -1,10 +1,7 @@
 /*
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 3 of the License, or
- * (at your option) any later version.
+ * This software is distributed under BSD 3-clause license (see LICENSE file).
  *
- * Written (W) 2013 Soumyajit De
+ * Authors: Soeren Sonnenburg, Pan Deng, Soumyajit De, Björn Esser
  */
 
 #include <shogun/lib/common.h>

cmake/FindMetaExamples.cmake

@@ -40,7 +40,6 @@ function(get_excluded_meta_examples)
 	IF(NOT HAVE_LAPACK)
 		LIST(APPEND EXCLUDED_META_EXAMPLES
 			regression/linear_ridge_regression.sg
-			clustering/gmm.sg
 			distance/mahalanobis.sg
 			)
 	ENDIF()

doc/cookbook/source/examples/binary_classifier/kernel_svm.rst → doc/cookbook/source/examples/binary/kernel_support_vector_machine.rst

@@ -26,27 +26,27 @@ Example
 
 Imagine we have files with training and test data. We create CDenseFeatures (here 64 bit floats aka RealFeatures) and :sgclass:`CBinaryLabels` as
 
-.. sgexample:: kernel_svm.sg:create_features
+.. sgexample:: kernel_support_vector_machine.sg:create_features
 
 In order to run :sgclass:`CLibSVM`, we need to initialize a kernel like :sgclass:`CGaussianKernel` with training features and some parameters like :math:`C` and epsilon i.e. residual convergence parameter which is optional.
 
-.. sgexample:: kernel_svm.sg:set_parameters
+.. sgexample:: kernel_support_vector_machine.sg:set_parameters
 
 We create an instance of the :sgclass:`CLibSVM` classifier by passing it regularization coefficient, kernel and labels.
 
-.. sgexample:: kernel_svm.sg:create_instance
+.. sgexample:: kernel_support_vector_machine.sg:create_instance
 
 Then we train and apply it to test data, which here gives :sgclass:`CBinaryLabels`.
 
-.. sgexample:: kernel_svm.sg:train_and_apply
+.. sgexample:: kernel_support_vector_machine.sg:train_and_apply
 
 We can extract :math:`\alpha` and :math:`b`.
 
-.. sgexample:: kernel_svm.sg:extract_weights_bias
+.. sgexample:: kernel_support_vector_machine.sg:extract_weights_bias
 
 Finally, we can evaluate test performance via e.g. :sgclass:`CAccuracyMeasure`.
 
-.. sgexample:: kernel_svm.sg:evaluate_accuracy
+.. sgexample:: kernel_support_vector_machine.sg:evaluate_accuracy
 
 ----------
 References

doc/cookbook/source/examples/binary_classifier/lda.rst → doc/cookbook/source/examples/binary/linear_discriminant_analysis.rst

@@ -24,23 +24,23 @@ Example
 
 We create CDenseFeatures (here 64 bit floats aka RealFeatures) and :sgclass:`CBinaryLabels` from files with training and test data.
 
-.. sgexample:: lda.sg:create_features
+.. sgexample:: linear_discriminant_analysis.sg:create_features
 
 We create an instance of the :sgclass:`CLDA` classifier and set features and labels. By default, Shogun automatically chooses the decomposition method based on :math:`{N<=D}` or :math:`{N>D}`.
 
-.. sgexample:: lda.sg:create_instance
+.. sgexample:: linear_discriminant_analysis.sg:create_instance
 
 Then we train and apply it to test data, which here gives :sgclass:`CBinaryLabels`.
 
-.. sgexample:: lda.sg:train_and_apply
+.. sgexample:: linear_discriminant_analysis.sg:train_and_apply
 
 We can extract weights :math:`{\bf w}`.
 
-.. sgexample:: lda.sg:extract_weights
+.. sgexample:: linear_discriminant_analysis.sg:extract_weights
 
 We can evaluate test performance via e.g. :sgclass:`CAccuracyMeasure`.
 
-.. sgexample:: lda.sg:evaluate_accuracy
+.. sgexample:: linear_discriminant_analysis.sg:evaluate_accuracy
 
 ----------
 References

doc/cookbook/source/examples/binary_classifier/linear_svm.rst → doc/cookbook/source/examples/binary/linear_support_vector_machine.rst

@@ -26,27 +26,27 @@ Example
 
 Imagine we have files with training and test data. We create CDenseFeatures (here 64 bit floats aka RealFeatures) and :sgclass:`CBinaryLabels` as
 
-.. sgexample:: linear_svm.sg:create_features
+.. sgexample:: linear_support_vector_machine.sg:create_features
 
 In order to run :sgclass:`CLibLinear`, we need to initialize some parameters like :math:`C` and epsilon which is the residual convergence parameter of the solver.
 
-.. sgexample:: linear_svm.sg:set_parameters
+.. sgexample:: linear_support_vector_machine.sg:set_parameters
 
 We create an instance of the :sgclass:`CLibLinear` classifier by passing it regularization coefficient, features and labels. We here set the solver type to L2 regularized classification. There are many other solver types in :sgclass:`CLibLinear` to choose from.
 
-.. sgexample:: linear_svm.sg:create_instance
+.. sgexample:: linear_support_vector_machine.sg:create_instance
 
 Then we train and apply it to test data, which here gives :sgclass:`CBinaryLabels`.
 
-.. sgexample:: linear_svm.sg:train_and_apply
+.. sgexample:: linear_support_vector_machine.sg:train_and_apply
 
 We can extract :math:`{\bf w}` and :math:`b`.
 
-.. sgexample:: linear_svm.sg:extract_weights_bias
+.. sgexample:: linear_support_vector_machine.sg:extract_weights_bias
 
 We can evaluate test performance via e.g. :sgclass:`CAccuracyMeasure`.
 
-.. sgexample:: linear_svm.sg:evaluate_accuracy
+.. sgexample:: linear_support_vector_machine.sg:evaluate_accuracy
 
 ----------
 References

doc/cookbook/source/examples/binary_classifier/multiple_kernel_learning.rst → doc/cookbook/source/examples/binary/multiple_kernel_learning.rst

@@ -10,7 +10,7 @@ Multiple kernel learning (MKL) is based on convex combinations of arbitrary kern
 
 where :math:`\beta_k > 0`, :math:`\sum_{k=1}^{K} \beta_k = 1`, :math:`K` is the number of sub-kernels, :math:`\bf{k}` is a combined kernel, :math:`{\bf k}_i` is an individual kernel and :math:`{x_i}_i` are the training data.
 
-Classification is done by using Support Vector Machines (SVM). See :doc:`linear_svm` for more details. Optimal :math:`\alpha` and :math:`b` for SVM and :math:`\beta` are determined via training.
+Classification is done by using Support Vector Machines (SVM). See :doc:`linear_support_vector_machine` for more details. Optimal :math:`\alpha` and :math:`b` for SVM and :math:`\beta` are determined via training.
 
 See :cite:`sonnenburg2006large` for more details.
 
@@ -20,35 +20,35 @@ Example
 
 Imagine we have files with training and test data. We create CDenseFeatures (here 64 bit floats aka RealFeatures) and :sgclass:`CBinaryLabels` as
 
-.. sgexample:: mkl.sg:create_features
+.. sgexample:: multiple_kernel_learning.sg:create_features
 
 Then we create indvidual kernels like :sgclass:`CPolyKernel` and :sgclass:`CGaussianKernel` which will be later combined in one :sgclass:`CCombinedKernel`.
 
-.. sgexample:: mkl.sg:create_kernel
+.. sgexample:: multiple_kernel_learning.sg:create_kernel
 
 We create an instance of :sgclass:`CCombinedKernel` and append the :sgclass:`CKernel` objects.
 
-.. sgexample:: mkl.sg:create_combined_train
+.. sgexample:: multiple_kernel_learning.sg:create_combined_train
 
 We create an object of :sgclass:`CMKLClassification`, provide the combined kernel and labels before training it.
 
-.. sgexample:: mkl.sg:train_mkl
+.. sgexample:: multiple_kernel_learning.sg:train_mkl
 
 After training, we can extract :math:`\beta`, SVM coefficients :math:`\alpha` and :math:`b`.
 
-.. sgexample:: mkl.sg:extract_weights
+.. sgexample:: multiple_kernel_learning.sg:extract_weights
 
 We update the :sgclass:`CCombinedKernel` object for testing data.
 
-.. sgexample:: mkl.sg:create_combined_test
+.. sgexample:: multiple_kernel_learning.sg:create_combined_test
 
 We set the updated kernel and predict :sgclass:`CBinaryLabels` for test data.
 
-.. sgexample:: mkl.sg:mkl_apply
+.. sgexample:: multiple_kernel_learning.sg:mkl_apply
 
 Finally, we can evaluate test performance via e.g. :sgclass:`CAccuracyMeasure`.
 
-.. sgexample:: mkl.sg:evaluate_accuracy
+.. sgexample:: multiple_kernel_learning.sg:evaluate_accuracy
 
 ----------
 References

doc/cookbook/source/examples/clustering/gmm.rst → doc/cookbook/source/examples/clustering/gaussian_mixture_models.rst

@@ -20,27 +20,27 @@ Example
 
 We start by creating CDenseFeatures (here 64 bit floats aka RealFeatures) as
 
-.. sgexample:: gmm.sg:create_features
+.. sgexample:: gaussian_mixture_models.sg:create_features
 
 We initialize :sgclass:`CGMM`, passing the desired number of mixture components.
 
-.. sgexample:: gmm.sg:create_gmm_instance
+.. sgexample:: gaussian_mixture_models.sg:create_gmm_instance
 
 We provide training features to the :sgclass:`CGMM` object, train it by using EM algorithm and sample data-points from the trained model.
 
-.. sgexample:: gmm.sg:train_sample
+.. sgexample:: gaussian_mixture_models.sg:train_sample
 
 We extract parameters like :math:`\pi`, :math:`\mu_i` and :math:`\Sigma_i` for any componenet from the trained model.
 
-.. sgexample:: gmm.sg:extract_params
+.. sgexample:: gaussian_mixture_models.sg:extract_params
 
 We obtain log likelihood of belonging to clusters and being generated by this model.
 
-.. sgexample:: gmm.sg:cluster_output
+.. sgexample:: gaussian_mixture_models.sg:cluster_output
 
 We can also use Split-Merge Expectation-Maximization algorithm :cite:`ueda2000smem` for training.
 
-.. sgexample:: gmm.sg:training_smem
+.. sgexample:: gaussian_mixture_models.sg:training_smem
 
 ----------
 References

doc/cookbook/source/examples/converter/ica_fast.rst → doc/cookbook/source/examples/converter/independent_component_analysis_fast.rst

@@ -15,19 +15,19 @@ Example
 Given a dataset which we assume consists of linearly mixed signals, we create CDenseFeatures
 (RealFeatures, here 64 bit float values).
 
-.. sgexample:: ica_fast.sg:create_features
+.. sgexample:: independent_component_analysis_fast.sg:create_features
 
 We create the :sgclass:`CFastICA` instance, and set its parameter for the iterative optimization.
 
-.. sgexample:: ica_fast.sg:set_parameters
+.. sgexample:: independent_component_analysis_fast.sg:set_parameters
 
 Then we apply ICA, which gives the unmixed signals.
 
-.. sgexample:: ica_fast.sg:apply_convert
+.. sgexample:: independent_component_analysis_fast.sg:apply_convert
 
 We can also extract the estimated mixing matrix.
 
-.. sgexample:: ica_fast.sg:extract
+.. sgexample:: independent_component_analysis_fast.sg:extract
 
 ----------
 References

doc/cookbook/source/examples/evaluation/cross_validation_mkl_weights_storage.rst → doc/cookbook/source/examples/evaluation/cross_validation_multiple_kernel_learning_weights_storage.rst

@@ -15,39 +15,39 @@ We'll use as example a classification problem solvable by using :sgclass:`CMKLCl
 For the sake of brevity, we'll skip the initialization of features, kernels and so on
 (see :doc:`../regression/multiple_kernel_learning` for a more complete example of MKL usage).
 
-.. sgexample:: cross_validation_mkl_weight_storage.sg:create_classifier
+.. sgexample:: cross_validation_multiple_kernel_learning_weights_storage.sg:create_classifier
 
 Firstly, we initialize a splitting strategy :sgclass:`CStratifiedCrossValidationSplitting`, which is needed
 to divide the dataset into folds, and an evaluation criterium :sgclass:`CAccuracyMeasure`, to evaluate the
 performance of the trained models. Secondly, we create the :sgclass:`CCrossValidation` instance.
 We set also the number of cross validation's runs.
 
-.. sgexample:: cross_validation_mkl_weight_storage.sg:create_cross_validation
+.. sgexample:: cross_validation_multiple_kernel_learning_weights_storage.sg:create_cross_validation
 
 To observe also the partial folds' results, we create a cross validation's observer :sgclass:`CParameterObserverCV`
 and then we register it into the :sgclass:`CCrossValidation` instance.
 
-.. sgexample:: cross_validation_mkl_weight_storage.sg:create_observer
+.. sgexample:: cross_validation_multiple_kernel_learning_weights_storage.sg:create_observer
 
 Finally, we evaluate the model and get the results (aka a :sgclass:`CCrossValidationResult` instance).
 
-.. sgexample:: cross_validation_mkl_weight_storage.sg:evaluate_and_get_result
+.. sgexample:: cross_validation_multiple_kernel_learning_weights_storage.sg:evaluate_and_get_result
 
 We get the :math:`mean` of all the evaluation results and its standard deviation :math:`stddev`.
 
-.. sgexample:: cross_validation_mkl_weight_storage.sg:get_results
+.. sgexample:: cross_validation_multiple_kernel_learning_weights_storage.sg:get_results
 
 We can get more information about the single cross validation's runs and folds by using the observer we set before, like the kernels' weights.
 We get the :sgclass:`CMKLClassification` machine used during the first run and trained on the first fold.
 
-.. sgexample:: cross_validation_mkl_weight_storage.sg:get_fold_machine
+.. sgexample:: cross_validation_multiple_kernel_learning_weights_storage.sg:get_fold_machine
 
 Then, from the trained machine, we get the weights :math:`\mathbf{w}` of its kernels.
 
-.. sgexample:: cross_validation_mkl_weight_storage.sg:get_weights
+.. sgexample:: cross_validation_multiple_kernel_learning_weights_storage.sg:get_weights
 
 ----------
 References
 ----------
 
-:wiki:`Cross-validation_(statistics)`
+:wiki:`Cross-validation_(statistics)`

doc/cookbook/source/examples/gaussian_processes/gaussian_process_classifier.rst → doc/cookbook/source/examples/gaussian_process/classifier.rst

@@ -4,7 +4,7 @@ Gaussian Process Classifier
 
 Application of Gaussian processes in binary and multi-class classification.
 `See Gaussian process regression cookbook
-<http://shogun.ml/cookbook/latest/examples/gaussian_processes/gaussian_process_regression.html>`_
+<http://shogun.ml/cookbook/latest/examples/gaussian_process/regression.html>`_
 and :cite:`Rasmussen2005GPM` for more information on Gaussian processes.
 
 -------
@@ -13,12 +13,12 @@ Example
 
 Imagine we have files with training and test data. We create CDenseFeatures (here 64 bit floats aka RealFeatures) and :sgclass:`CMulticlassLabels` as
 
-.. sgexample:: gaussian_process_classifier.sg:create_features
+.. sgexample:: classifier.sg:create_features
 
 To fit the input (training) data :math:`\mathbf{X}`, we have to choose appropriate :sgclass:`CMeanFunction`
 and  :sgclass:`CKernel`. Here we use a basic :sgclass:`CConstMean` and a :sgclass:`CGaussianKernel` with chosen width parameter.
 
-.. sgexample:: gaussian_process_classifier.sg:create_appropriate_kernel_and_mean_function
+.. sgexample:: classifier.sg:create_appropriate_kernel_and_mean_function
 
 We need to specify the inference method to find the posterior distribution of the function values :math:`\mathbf{f}`.
 Here we choose to perform Laplace approximation inference method with an instance of :sgclass:`CMultiLaplaceInferenceMethod` (See Chapter 18.2 in :cite:`barber2012bayesian` for a detailed introduction)
@@ -27,20 +27,20 @@ the training features, the mean function, the labels and an instance of :sgclass
 to specify the distribution of the targets/labels as above.
 Finally we create an instance of the :sgclass:`CGaussianProcessClassification` classifier.
 
-.. sgexample:: gaussian_process_classifier.sg:create_instance
+.. sgexample:: classifier.sg:create_instance
 
 Then we can train the model and evaluate the predictive distribution.
 We get predicted :sgclass:`CMulticlassLabels`.
 
-.. sgexample:: gaussian_process_classifier.sg:train_and_apply
+.. sgexample:: classifier.sg:train_and_apply
 
 We can extract the probabilities:
 
-.. sgexample:: gaussian_process_classifier.sg:extract_the_probabilities
+.. sgexample:: classifier.sg:extract_the_probabilities
 
 We can evaluate test performance via e.g. :sgclass:`CMulticlassAccuracy`.
 
-.. sgexample:: gaussian_process_classifier.sg:evaluate_accuracy
+.. sgexample:: classifier.sg:evaluate_accuracy
 
 ----------
 References

doc/cookbook/source/examples/gaussian_processes/gaussian_process_regression.rst → doc/cookbook/source/examples/gaussian_process/regression.rst

@@ -24,32 +24,32 @@ Example
 
 Imagine we have files with training and test data. We create `CDenseFeatures` (here 64 bit floats aka RealFeatures) and :sgclass:`CRegressionLabels` as:
 
-.. sgexample:: gaussian_process_regression.sg:create_features
+.. sgexample:: regression.sg:create_features
 
 To fit the input (training) data :math:`\mathbf{X}`, we have to choose an appropriate :sgclass:`CMeanFunction` and  :sgclass:`CKernel` and instantiate them. Here we use a basic :sgclass:`CZeroMean` and a :sgclass:`CGaussianKernel` with chosen width parameter.
 
-.. sgexample:: gaussian_process_regression.sg:create_appropriate_kernel_and_mean_function
+.. sgexample:: regression.sg:create_appropriate_kernel_and_mean_function
 
 We need to specify the inference method to find the posterior distribution of the function values :math:`\mathbf{f}`. Here we choose to perform exact inference with an instance of :sgclass:`CExactInferenceMethod` and pass it the chosen kernel, the training features, the mean function, the labels and an instance of :sgclass:`CGaussianLikelihood`, to specify the distribution of the targets/labels as above. Finally we generate a CGaussianProcessRegression class to be trained.
 
-.. sgexample:: gaussian_process_regression.sg:create_instance
+.. sgexample:: regression.sg:create_instance
 
 Then we can train the model and evaluate the predictive distribution. We get predicted :sgclass:`CRegressionLabels`.
 
-.. sgexample:: gaussian_process_regression.sg:train_and_apply
+.. sgexample:: regression.sg:train_and_apply
 
 We can compute the predictive variances as
 
-.. sgexample:: gaussian_process_regression.sg:compute_variance
+.. sgexample:: regression.sg:compute_variance
 
 The prediction above is based on arbitrarily set hyperparameters :math:`\boldsymbol{\theta}`: kernel width :math:`\tau`, kernel scaling :math:`\gamma` and observation noise :math:`\sigma^2`. We can also learn these parameters by optimizing the marginal likelihood :math:`p(\mathbf{y}|\mathbf{X}, \boldsymbol{\theta})` w.r.t. :math:`\boldsymbol{\theta}`.
 To do this, we define a :sgclass:`CGradientModelSelection`, passing to it a :sgclass:`CGradientEvaluation` with its own :sgclass:`CGradientCriterion`, specifying the gradient scheme and direction. Then we can follow the gradient and apply the chosen :math:`\boldsymbol{\theta}` back to the CGaussianProcessRegression instance.
 
-.. sgexample:: gaussian_process_regression.sg:optimize_marginal_likelihood
+.. sgexample:: regression.sg:optimize_marginal_likelihood
 
 Finally, we evaluate the :sgclass:`CMeanSquaredError` and the (negative log) marginal likelihood for the optimized hyperparameters.
 
-.. sgexample:: gaussian_process_regression.sg:evaluate_error_and_marginal_likelihood
+.. sgexample:: regression.sg:evaluate_error_and_marginal_likelihood
 
 ----------
 References