Fn_sparse_random_ecoc

Functions
imat	ecocpak::create_sparse_matrix_from_rand (const double *random_matrix, const u32 n_rows, const u32 n_cols)
bool	ecocpak::identical_columns_sparse (const imat &coding_matrix)
imat	ecocpak::create_sparse_random_matrix (const u32 n_classes, const u32 n_classifiers, const u32 n_matrices)
u32	ecocpak::sparse_random_ecoc (const mat &training_samples, const icolvec &training_labels, const mat &testing_samples, const icolvec &testing_labels, const int decoding_strategy, const int classifiers_type, const u32 n_matrices, const u32 n_desired_classifiers, const bool verbose, ofstream &verbose_output, double &execution_time)

---

Function Documentation

imat ecocpak::create_sparse_matrix_from_rand	(	const double *	random_matrix,
		const u32	n_rows,
		const u32	n_cols
	)

• Creates sparse random ECOC matrix from a random real valued matrix with its elements in [0,1].

• Auxiliary function for create_sparse_random_matrix().

• Input Arguments:
  • random_matrix : A pointer (i.e., address) to a random produced matrix.
  • n_rows : Number of input matrix rows.
  • n_cols : Number of input matrix columns.

• Output Arguments:
  • Void.

• Return Argument:
  • Returns a sparse random coding matrix.

imat ecocpak::create_sparse_random_matrix	(	const u32	n_classes,
		const u32	n_classifiers,
		const u32	n_matrices
	)

• Creates sparse random ECOC coding matrix.

• Input Arguments:
  • n_classes : Number of classes.
  • n_classifiers : Number of desired classifiers (i.e., number of coding matrix columns).
  • n_matrices : Number of valid dense random coding matrices to be produced.

• Output Arguments:
  • Void.

• Return Argument:
  • Returns a valid sparse random coding matrix with integral elements.

bool ecocpak::identical_columns_sparse

(

const imat &

coding_matrix

)

• Check wether input coding matrix has identical columns (i.e., the same problem is examined)

• Auxiliary function for create_sparse_random_matrix().

• Input Arguments:
  • coding_matrix : Input coding matrix.

• Output Arguments:
  • Void.

• Return Argument:
  • Returns true if input coding matrix has identical columns, false otherwise.

u32 ecocpak::sparse_random_ecoc	(	const mat &	training_samples,
		const icolvec &	training_labels,
		const mat &	testing_samples,
		const icolvec &	testing_labels,
		const int	decoding_strategy,
		const int	classifiers_type,
		const u32	n_matrices,
		const u32	n_desired_classifiers,
		const bool	verbose,
		ofstream &	verbose_output,
		double &	execution_time
	)

• Error Correcting Output Codes with sparse random coding, training and testing procedure

• Input Arguments:
  • training_samples : A 2D input matrix where its rows represent training sample vectors.
  • training_labels : The respective class labels of training_samples input matrix.
  • testing_samples : A 2D input matrix where its rows represents testing sample vectors.
  • testing_labels : The respective class labels of testing_samples input matrix.
  • decoding_strategy : User's option specifying the type of decoding methode that will be used.
  • classifiers_type : Type of classifier.
  • n_desired_classifiers : Number of desired classifiers (i.e., number of coding matrix columns).
  • n_matrices : Number of valid dense random coding matrices to be produced.

• Output Arguments:
  • coding_matrix : Coding matrix.
  • classifiers_vector : Vector of classifiers objects.
  • execution_time : Execution time.

• Return Arguments:
  • Number of misclassified test samples.