lti::supervisedSequenceClassifier Class Reference

Abstract class, parent of all supervised classifiers for sequences. More...

#include <ltiSupervisedSequenceClassifier.h>

Inheritance diagram for lti::supervisedSequenceClassifier:

[legend]

Collaboration diagram for lti::supervisedSequenceClassifier:

[legend]

List of all members.

Public Member Functions
	supervisedSequenceClassifier ()
	supervisedSequenceClassifier (const supervisedSequenceClassifier &other)
virtual const char *	getTypeName () const
supervisedSequenceClassifier &	copy (const supervisedSequenceClassifier &other)
supervisedSequenceClassifier &	operator= (const supervisedSequenceClassifier &other)
const parameters &	getParameters () const
virtual bool	train (const std::vector< sequence< dvector > > &input, const ivector &ids)=0
virtual bool	classify (const sequence< dvector > &observations, outputVector &result)=0
Protected Member Functions
bool	makeOutputTemplate (const int &outSize, const std::vector< sequence< dvector > > &data, const ivector &ids)

---

Detailed Description

Abstract class, parent of all supervised classifiers for sequences.

This class defines the interface for training methods of classifiers that work on time series.

---

Constructor & Destructor Documentation

lti::supervisedSequenceClassifier::supervisedSequenceClassifier

(

)

default constructor

lti::supervisedSequenceClassifier::supervisedSequenceClassifier

(

const supervisedSequenceClassifier &

other

)

copy constructor

Parameters:

other

the object to be copied

---

Member Function Documentation

virtual bool lti::supervisedSequenceClassifier::classify	(	const sequence< dvector > &	observations,
		outputVector &	result
	)			[pure virtual]

Classification.

Classifies the sequence of features and returns the outputVector with the classification result. Since sequence classifiers usually keep an inner state, this method is not const.

Parameters:

	observations	the sequence to be classified
	result	the result of the classification

Returns:

false if an error occurred during classification else true

Implemented in lti::hmmClassifier.

supervisedSequenceClassifier& lti::supervisedSequenceClassifier::copy

(

const supervisedSequenceClassifier &

other

)

copy data of "other" functor.

Parameters:

other

the functor to be copied

Returns:

a reference to this functor object

Reimplemented from lti::classifier.

Reimplemented in lti::hmmClassifier, and lti::hmmOnlineClassifier.

Referenced by operator=().

const parameters& lti::supervisedSequenceClassifier::getParameters

(

)

const

returns used parameters

Reimplemented from lti::classifier.

Reimplemented in lti::hmmClassifier, and lti::hmmOnlineClassifier.

virtual const char* lti::supervisedSequenceClassifier::getTypeName

(

)

const [virtual]

returns the name of this type ("supervisedSequenceClassifier")

Reimplemented from lti::classifier.

Reimplemented in lti::hmmClassifier, and lti::hmmOnlineClassifier.

bool lti::supervisedSequenceClassifier::makeOutputTemplate	(	const int &	outSize,
		const std::vector< sequence< dvector > > &	data,
		const ivector &	ids
	)			[protected]

Sets the outputTemplate probability distributions according to the classification of the given data.

The distributions are built by the follwing rule:

1. Classify next data sequence
2. For the position in the output with the highest probability increase the count for the actual id by one.
3. While there is more data go back to 1
4. For each position: divide each count by total number of counts

This results in a distribution over the ids that caused highest probability for each position of the output.

Parameters:

	outSize	size of the outputTemplate
	data	train of validation data sequences
	ids	ids of the data-vectors

Returns:

false upon error

supervisedSequenceClassifier& lti::supervisedSequenceClassifier::operator=

(

const supervisedSequenceClassifier &

other

)

[inline]

Alias for "copy".

Parameters:

other

the functor to be copied

Returns:

a reference to this functor object

Reimplemented from lti::classifier.

Reimplemented in lti::hmmClassifier, and lti::hmmOnlineClassifier.

References copy().

virtual bool lti::supervisedSequenceClassifier::train	(	const std::vector< sequence< dvector > > &	input,
		const ivector &	ids
	)			[pure virtual]

Supervised sequence training.

The sequences in the input vector are used for training the sequence classifier. Each sequence is associated with an id given in the vector ids.

Parameters:

	input	vector of training sequences
	ids	the classes of the training sequences

Returns:

true if successful, false otherwise. (if false you can check the error message with getStatusString())

Implemented in lti::hmmClassifier.

---

The documentation for this class was generated from the following file:

• ltiSupervisedSequenceClassifier.h