nb_freq.c File Reference

Naive Bayes non-parametric marker frequency 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 "xml.h"
#include "haplo_groups.h"
#include "nb_freq.h"

Go to the source code of this file.

Functions
void	train_nb_freq_model (NB_freq_model **model_out, const Vector_u32 *labels, const Matrix_i32 *markers, const Vector_d *label_priors)
	Trains a Naive Bayes non-parametric marker frequency model.
Error *	predict_label_with_nb_freq_model (uint32_t *label_out, double *confidence_out, const Vector_i32 *markers, const NB_freq_model *model, uint32_t order)
	Predicts the label for a marker sample using a Naive Bayes non-parametric marker frequency model.
Error *	predict_labels_with_nb_freq_model (Vector_u32 **labels_out, Vector_d **confidence_out, const Matrix_i32 *markers, const NB_freq_model *model)
	Predicts the labels for a set of marker samples using a Naive Bayes non-parametric marker frequency model.
Error *	read_nb_freq_model (NB_freq_model **model_out, const char *fname)
	Reads a Naive Bayes non-parametric marker frequency model.
Error *	write_nb_freq_model (NB_freq_model *model, const char *fname)
	Writes a Naive Bayes non-parametric marker frequency model.
void	free_nb_freq_model (NB_freq_model *model)
	Frees a Naive Bayes non-parametric marker frequency model.
static Error *	read_nb_freq_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_freq_model_tree_from_xml_node (NB_freq_model_tree **tree_out, const NB_freq_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_freq_model_node (NB_freq_model_node *node, const Vector_u32 *labels, const Matrix_i32 *markers)
	Trains a model in the tree using node-specific data.
Error *	train_nb_freq_model_tree (NB_freq_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 non-parametric marker frequency model tree .
static Error *	recursively_predict_label_in_model_tree (uint32_t *label_out, double *confidence_out, const NB_freq_model_tree *tree, const Vector_i32 *markers_v, uint32_t order)
	Recursively predicts a label from a model tree.
Error *	predict_labels_with_nb_freq_model_tree (Vector_u32 **labels_out, Vector_d **confidence_out, const Matrix_i32 *markers, const NB_freq_model_tree *tree, uint32_t order)
	Predicts the labels for a set of marker samples using a Naive Bayes non-parametric marker frequency model tree.
static Error *	read_nb_freq_model_node (NB_freq_model_node *node, const char *model_dirname)
	Recursively reads the model in a node from its file name.
Error *	read_nb_freq_model_tree (NB_freq_model_tree **tree_out, const char *tree_xml_fname, const char *tree_dtd_fname, const char *model_dirname)
	Reads a Naive Bayes non-parametric marker frequency model tree from.
Error *	write_nb_freq_model_tree (const NB_freq_model_tree *tree, const char *model_dirname)
	Writes a Naive Bayes non-parametric marker frequency model tree.
void	free_nb_freq_model_tree (NB_freq_model_tree *tree)
	Frees a Naive Bayes non-parametric marker frequency model tree.

---

Detailed Description

Naive Bayes non-parametric marker frequency classifier.

Author:

Joseph Schlecht

License:

Creative Commons BY-NC-SA 3.0

Definition in file nb_freq.c.

---

Function Documentation

void train_nb_freq_model	(	NB_freq_model **	model_out,
		const Vector_u32 *	labels,
		const Matrix_i32 *	markers,
		const Vector_d *	label_priors
	)

Trains a Naive Bayes non-parametric marker frequency 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. Set to NULL to used priors calculated from frequences in the training data.

Definition at line 90 of file nb_freq.c.

Error* predict_label_with_nb_freq_model	(	uint32_t *	label_out,
		double *	confidence_out,
		const Vector_i32 *	markers,
		const NB_freq_model *	model,
		uint32_t	order
	)

Predicts the label for a marker sample using a Naive Bayes non-parametric marker frequency 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 269 of file nb_freq.c.

Error* predict_labels_with_nb_freq_model	(	Vector_u32 **	labels_out,
		Vector_d **	confidence_out,
		const Matrix_i32 *	markers,
		const NB_freq_model *	model
	)

Predicts the labels for a set of marker samples using a Naive Bayes non-parametric marker frequency 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 376 of file nb_freq.c.

Error* read_nb_freq_model	(	NB_freq_model **	model_out,
		const char *	fname
	)

Reads a Naive Bayes non-parametric marker frequency 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 433 of file nb_freq.c.

Error* write_nb_freq_model	(	NB_freq_model *	model,
		const char *	fname
	)

Writes a Naive Bayes non-parametric marker frequency 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 522 of file nb_freq.c.

void free_nb_freq_model

(

NB_freq_model *

model

)

Frees a Naive Bayes non-parametric marker frequency model.

Parameters:

model

Model to free.

Definition at line 594 of file nb_freq.c.

static Error* read_nb_freq_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 607 of file nb_freq.c.

static Error* create_nb_freq_model_tree_from_xml_node	(	NB_freq_model_tree **	tree_out,
		const NB_freq_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 660 of file nb_freq.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 803 of file nb_freq.c.

static Error* train_nb_freq_model_node	(	NB_freq_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 874 of file nb_freq.c.

Error* train_nb_freq_model_tree	(	NB_freq_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 non-parametric marker frequency 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 924 of file nb_freq.c.

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

Recursively predicts a label from a model tree.

Definition at line 974 of file nb_freq.c.

Error* predict_labels_with_nb_freq_model_tree	(	Vector_u32 **	labels_out,
		Vector_d **	confidence_out,
		const Matrix_i32 *	markers,
		const NB_freq_model_tree *	tree,
		uint32_t	order
	)

Predicts the labels for a set of marker samples using a Naive Bayes non-parametric marker frequency 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 1032 of file nb_freq.c.

static Error* read_nb_freq_model_node	(	NB_freq_model_node *	node,
		const char *	model_dirname
	)		[static]

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

Definition at line 1084 of file nb_freq.c.

Error* read_nb_freq_model_tree	(	NB_freq_model_tree **	tree_out,
		const char *	tree_xml_fname,
		const char *	tree_dtd_fname,
		const char *	model_dirname
	)

Reads a Naive Bayes non-parametric marker frequency 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 1122 of file nb_freq.c.

Error* write_nb_freq_model_tree	(	const NB_freq_model_tree *	tree,
		const char *	model_dirname
	)

Writes a Naive Bayes non-parametric marker frequency model tree.

Parameters:

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

Definition at line 1174 of file nb_freq.c.

void free_nb_freq_model_tree

(

NB_freq_model_tree *

tree

)

Frees a Naive Bayes non-parametric marker frequency model tree.

Parameters:

tree	Model tree to free.

Definition at line 1209 of file nb_freq.c.