nb_gmm.c File Reference

Naive Bayes Gaussian mixture model classifier. More...

#include <config.h>
#include <stdlib.h>
#include <stdio.h>
#include <inttypes.h>
#include <assert.h>
#include <string.h>
#include <errno.h>
#include <math.h>
#include <libxml/tree.h>
#include <libxml/parser.h>
#include <libxml/valid.h>
#include <jwsc/base/error.h>
#include <jwsc/base/limits.h>
#include <jwsc/base/file_io.h>
#include <jwsc/vector/vector.h>
#include <jwsc/vector/vector_io.h>
#include <jwsc/vector/vector_math.h>
#include <jwsc/matrix/matrix.h>
#include <jwsc/matrix/matrix_io.h>
#include <jwsc/matblock/matblock.h>
#include <jwsc/matblock/matblock_io.h>
#include <jwsc/stat/gmm.h>
#include "xml.h"
#include "haplo_groups.h"
#include "nb_gmm.h"

Go to the source code of this file.

Functions
void	train_nb_gmm_model (NB_gmm_model **model_out, const Vector_u32 *labels, const Matrix_i32 *markers, const Vector_d *label_priors, uint32_t num_components)
	Trains a Naive Bayes Gaussian mixture model.
Error *	predict_label_with_nb_gmm_model (uint32_t *label_out, double *confidence_out, const Vector_i32 *markers, const NB_gmm_model *model, uint32_t order)
	Predicts the label for a marker sample using a Naive Bayes Gaussian mixture model.
Error *	predict_labels_with_nb_gmm_model (Vector_u32 **labels_out, Vector_d **confidence_out, const Matrix_i32 *markers, const NB_gmm_model *model)
	Predicts the labels for a set of marker samples using a Naive Bayes Gaussian mixture model.
Error *	read_nb_gmm_model (NB_gmm_model **model_out, const char *fname)
	Reads a Naive Bayes Gaussian mixture model.
Error *	write_nb_gmm_model (NB_gmm_model *model, const char *fname)
	Writes a Naive Bayes Gaussian mixture model.
void	free_nb_gmm_model (NB_gmm_model *model)
	Frees a Naive Bayes Gaussian mixture model.
static Error *	read_nb_gmm_xml_doc (xmlDoc **xml_doc_out, const char *xml_fname, const char *dtd_fname)
	Reads and and optionally validates an XML document.
static Error *	create_nb_gmm_model_tree_from_xml_node (NB_gmm_model_tree **tree_out, const NB_gmm_model_tree *parent, uint32_t parent_label, xmlNode *xml_node)
	Allocates and initializes a model tree from an xml node.
static Error *	create_model_training_data (Vector_u32 **train_labels_out, Matrix_i32 **train_markers_out, const Vector_u32 *data_labels, const Matrix_i32 *data_markers, const Vector_u32 *model_labels, Vector_u32 *const *model_altlabels)
	Creates a set of training data as a model-specific labeled subset of a larger data set.
static Error *	train_nb_gmm_model_node (NB_gmm_model_node *node, const Vector_u32 *labels, const Matrix_i32 *markers)
	Trains a model in the tree using node-specific data.
Error *	train_nb_gmm_model_tree (NB_gmm_model_tree **tree_out, const Vector_u32 *labels, const Matrix_i32 *markers, const char *tree_xml_fname, const char *tree_dtd_fname)
	Trains a Naive Bayes Gaussian mixture model tree .
static Error *	recursively_predict_label_in_model_tree (uint32_t *label_out, double *confidence_out, const NB_gmm_model_tree *tree, const Vector_i32 *markers_v, uint32_t order)
	Recursively predicts a label from a model tree.
Error *	predict_labels_with_nb_gmm_model_tree (Vector_u32 **labels_out, Vector_d **confidence_out, const Matrix_i32 *markers, const NB_gmm_model_tree *tree, uint32_t order)
	Predicts the labels for a set of marker samples using a Naive Bayes Gaussian mixture model tree.
static Error *	read_nb_gmm_model_node (NB_gmm_model_node *node, const char *model_dirname)
	Recursively reads the model in a node from its file name.
Error *	read_nb_gmm_model_tree (NB_gmm_model_tree **tree_out, const char *tree_xml_fname, const char *tree_dtd_fname, const char *model_dirname)
	Reads a Naive Bayes Gaussian mixture model tree from.
Error *	write_nb_gmm_model_tree (const NB_gmm_model_tree *tree, const char *model_dirname)
	Writes a Naive Bayes Gaussian mixture model tree.
void	free_nb_gmm_model_tree (NB_gmm_model_tree *tree)
	Frees a Naive Bayes Gaussian mixture model tree.

---

Detailed Description

Naive Bayes Gaussian mixture model classifier.

Author:

Joseph Schlecht

License:

Creative Commons BY-NC-SA 3.0

Definition in file nb_gmm.c.

---

Function Documentation

void train_nb_gmm_model	(	NB_gmm_model **	model_out,
		const Vector_u32 *	labels,
		const Matrix_i32 *	markers,
		const Vector_d *	label_priors,
		uint32_t	num_components
	)

Trains a Naive Bayes Gaussian mixture model.

Parameters:

model_out	Result parameter. If *model_out is NULL, a model is allocated; otherwise its space is re-used.
labels	Labels for training, with ith element as corresponding to the ith sample in markers.
markers	Markers for training, with ith row as a sample corresponding to the ith label in labels.
label_priors	Priors for each of the labels.
num_components	Number of mixture components.

Definition at line 92 of file nb_gmm.c.

Error* predict_label_with_nb_gmm_model	(	uint32_t *	label_out,
		double *	confidence_out,
		const Vector_i32 *	markers,
		const NB_gmm_model *	model,
		uint32_t	order
	)

Predicts the label for a marker sample using a Naive Bayes Gaussian mixture model.

Parameters:

label_out	Result parameter.
confidence_out	Result parameter.
markers	Marker data to predict.
model	Trained model to use for predicting.
order	Indicates which of the The kth best prediction values to return. Starts with 0 representing the overall best, 1 the next best and so on.

Returns:

On success, NULL is returned; otherwise an error is returned.

Definition at line 215 of file nb_gmm.c.

Error* predict_labels_with_nb_gmm_model	(	Vector_u32 **	labels_out,
		Vector_d **	confidence_out,
		const Matrix_i32 *	markers,
		const NB_gmm_model *	model
	)

Predicts the labels for a set of marker samples using a Naive Bayes Gaussian mixture model.

Parameters:

labels_out	Result parameter. If *labels_out is NULL, it is allocated; otherwise its space is re-used.
confidence_out	Result parameter. If *confidence_out is NULL, it is allocated; otherwise its space is re-used.
markers	Marker data to predict. Each row is a sample for prediction, corresponding to an element in the result parameters.
model	Trained model to use for predicting.

Returns:

On success, NULL is returned; otherwise an error is returned, but the result parameters are not freed.

Definition at line 338 of file nb_gmm.c.

Error* read_nb_gmm_model	(	NB_gmm_model **	model_out,
		const char *	fname
	)

Reads a Naive Bayes Gaussian mixture model.

Parameters:

model_out	Result parameter. If *model_out is NULL, a model is allocated; otherwise its space is re-used.
fname	File to read the model from.

Returns:

On success, NULL is returned; otherwise an error is returned but does not free result parameter.

Definition at line 395 of file nb_gmm.c.

Error* write_nb_gmm_model	(	NB_gmm_model *	model,
		const char *	fname
	)

Writes a Naive Bayes Gaussian mixture model.

Parameters:

model	Model to write.
fname	File to write to.

Returns:

On success, NULL is returned; otherwise an error is returned.

Definition at line 491 of file nb_gmm.c.

void free_nb_gmm_model

(

NB_gmm_model *

model

)

Frees a Naive Bayes Gaussian mixture model.

Parameters:

model

Model to free.

Definition at line 542 of file nb_gmm.c.

static Error* read_nb_gmm_xml_doc	(	xmlDoc **	xml_doc_out,
		const char *	xml_fname,
		const char *	dtd_fname
	)		[static]

Reads and and optionally validates an XML document.

Definition at line 577 of file nb_gmm.c.

static Error* create_nb_gmm_model_tree_from_xml_node	(	NB_gmm_model_tree **	tree_out,
		const NB_gmm_model_tree *	parent,
		uint32_t	parent_label,
		xmlNode *	xml_node
	)		[static]

Allocates and initializes a model tree from an xml node.

Does not train or read the model; leaves the model field as NULL in each tree node.

Definition at line 630 of file nb_gmm.c.

static Error* create_model_training_data	(	Vector_u32 **	train_labels_out,
		Matrix_i32 **	train_markers_out,
		const Vector_u32 *	data_labels,
		const Matrix_i32 *	data_markers,
		const Vector_u32 *	model_labels,
		Vector_u32 *const *	model_altlabels
	)		[static]

Creates a set of training data as a model-specific labeled subset of a larger data set.

Definition at line 783 of file nb_gmm.c.

static Error* train_nb_gmm_model_node	(	NB_gmm_model_node *	node,
		const Vector_u32 *	labels,
		const Matrix_i32 *	markers
	)		[static]

Trains a model in the tree using node-specific data.

Definition at line 854 of file nb_gmm.c.

Error* train_nb_gmm_model_tree	(	NB_gmm_model_tree **	tree_out,
		const Vector_u32 *	labels,
		const Matrix_i32 *	markers,
		const char *	tree_xml_fname,
		const char *	tree_dtd_fname
	)

Trains a Naive Bayes Gaussian mixture model tree .

Parameters:

tree_out	Result parameter.
labels	Sample group labels.
markers	Sample marker values.
tree_xml_fname	XML file containing the model tree information.
tree_dtd_fname	DTD file for validating the XML file, can be NULL.

Definition at line 904 of file nb_gmm.c.

static Error* recursively_predict_label_in_model_tree	(	uint32_t *	label_out,
		double *	confidence_out,
		const NB_gmm_model_tree *	tree,
		const Vector_i32 *	markers_v,
		uint32_t	order
	)		[static]

Recursively predicts a label from a model tree.

Definition at line 954 of file nb_gmm.c.

Error* predict_labels_with_nb_gmm_model_tree	(	Vector_u32 **	labels_out,
		Vector_d **	confidence_out,
		const Matrix_i32 *	markers,
		const NB_gmm_model_tree *	tree,
		uint32_t	order
	)

Predicts the labels for a set of marker samples using a Naive Bayes Gaussian mixture model tree.

Parameters:

labels_out	Result parameter. If *labels_out is NULL, it is allocated; otherwise its space is re-used.
confidence_out	Result parameter. If *confidence_out is NULL, it is allocated; otherwise its space is re-used.
markers	Marker data to predict. Each row is a sample for prediction, corresponding to an element in the result parameters.
tree	Trained model tree to use for predicting.
order	The kth order prediction value. Zero being the best prediction.

Returns:

On success, NULL is returned; otherwise an error is returned, but the result parameters are not freed.

Definition at line 1012 of file nb_gmm.c.

static Error* read_nb_gmm_model_node	(	NB_gmm_model_node *	node,
		const char *	model_dirname
	)		[static]

Recursively reads the model in a node from its file name.

Definition at line 1064 of file nb_gmm.c.

Error* read_nb_gmm_model_tree	(	NB_gmm_model_tree **	tree_out,
		const char *	tree_xml_fname,
		const char *	tree_dtd_fname,
		const char *	model_dirname
	)

Reads a Naive Bayes Gaussian mixture model tree from.

Parameters:

tree_out	Result parameter.
tree_xml_fname	XML file containing the model tree information.
tree_dtd_fname	DTD file for validating the XML file, can be NULL.
model_dirname	Directory prefixing each model file in the tree.

Definition at line 1102 of file nb_gmm.c.

Error* write_nb_gmm_model_tree	(	const NB_gmm_model_tree *	tree,
		const char *	model_dirname
	)

Writes a Naive Bayes Gaussian mixture model tree.

Parameters:

tree	Model tree to write.
model_dirname	Directory to prefix to each model file in the tree.

Definition at line 1154 of file nb_gmm.c.

void free_nb_gmm_model_tree

(

NB_gmm_model_tree *

tree

)

Frees a Naive Bayes Gaussian mixture model tree.

Parameters:

tree	Model tree to free.

Definition at line 1189 of file nb_gmm.c.