latest version v1.9 - last update 10 Apr 2010 |
class which implements RBF networks. More...
#include <ltiRbf.h>
Classes | |
class | initFunctor1 |
initialization functor for rbf More... | |
class | initFunctor2 |
initialization functor for rbf networks More... | |
class | layer1 |
layer 1 for rbf networks More... | |
class | layer2 |
layer 2 for rbf networks More... | |
class | parameters |
parameters class for the RBF-Networks More... | |
Public Member Functions | |
rbf () | |
rbf (const rbf &other) | |
rbf (const parameters ¶m) | |
~rbf () | |
const char * | getTypeName () const |
bool | train (const dmatrix &input, const ivector &ids) |
bool | classify (const lti::dvector &feature, outputVector &result) const |
int | getDimOutputLayer () const |
void | setTestSet (const dmatrix &input, const ivector &ids) |
void | setTrainSet2 (const dmatrix &input, const ivector &ids) |
const parameters & | getParameters () const |
rbf & | copy (const rbf &other) |
rbf & | operator= (const rbf &other) |
virtual classifier * | clone () const |
Methods to access inner data. | |
Should not be used for anything but Axioms or others JNI interface to use this class. | |
layer1 | rbf1 |
layer2 | rbf2 |
int | sizeIn |
int | sizeOut |
ivector | trainID |
ivector | internToReal |
std::map< int, int > | realToIntern |
dmatrix | features |
dmatrix | train2Features |
ivector | train2Ids |
dvector | sigmas |
dmatrix | testFeatures |
ivector | testIDs |
ivector | trainStatIDs |
stdLayer::distancePropFunctor * | distProp |
stdLayer::dotPropFunctor * | dotProp |
stdLayer::gaussActFunctor * | gaussAct |
stdLayer::sigmoidActFunctor * | sigmoidAct |
std::map< int, dvector > | targetVec |
const dmatrix & | getWeights1 () |
const dvector & | getSigmas1 () |
const dmatrix & | getWeights2 () |
const ivector & | getIds () |
void | setSigmas (const dvector &s) |
void | setLayer1Data (const dmatrix &weights) |
void | setLayer2Data (const dmatrix &weights, const ivector &ids) |
void | initNet (const int &inSize, const int &outSize) |
virtual bool | write (ioHandler &handler, const bool complete=true) const |
virtual bool | read (ioHandler &handler, const bool complete=true) |
bool | calcObjProbs (ivector &ids, dmatrix &feats, stdLayer::distancePropFunctor &distProp) |
class which implements RBF networks.
lti::rbf::rbf | ( | ) |
Constructor.
lti::rbf::rbf | ( | const rbf & | other | ) |
Copy constructor.
lti::rbf::rbf | ( | const parameters & | param | ) |
Create with the given parameters.
lti::rbf::~rbf | ( | ) |
Destructor.
bool lti::rbf::calcObjProbs | ( | ivector & | ids, | |
dmatrix & | feats, | |||
stdLayer::distancePropFunctor & | distProp | |||
) | [protected] |
Calculate object probabilities with given features.
bool lti::rbf::classify | ( | const lti::dvector & | feature, | |
outputVector & | result | |||
) | const [virtual] |
Propagates a feature vector through the network.
The winner is set to the unit with the maximum value. If an error occurs, the outputVectors recect flag is set to true and false is returned.
Implements lti::supervisedInstanceClassifier.
virtual classifier* lti::rbf::clone | ( | ) | const [virtual] |
returns a pointer to a clone of this functor.
Implements lti::classifier.
copy data of "other" functor.
other | the functor to be copied |
Reimplemented from lti::supervisedInstanceClassifier.
Referenced by operator=().
int lti::rbf::getDimOutputLayer | ( | ) | const [inline] |
Get dimensions of the output layer.
References sizeOut.
const ivector& lti::rbf::getIds | ( | ) |
Returns the IDs of each neuron of the network.
const parameters& lti::rbf::getParameters | ( | ) | const |
returns used parameters
Reimplemented from lti::supervisedInstanceClassifier.
const char* lti::rbf::getTypeName | ( | ) | const [virtual] |
get type name
Reimplemented from lti::supervisedInstanceClassifier.
void lti::rbf::initNet | ( | const int & | inSize, | |
const int & | outSize | |||
) |
Set the size and init functors after data is transmitted to layers and sigmas are set.
Alias for copy.
other | the functor to be copied |
Reimplemented from lti::supervisedInstanceClassifier.
References copy().
virtual bool lti::rbf::read | ( | ioHandler & | handler, | |
const bool | complete = true | |||
) | [virtual] |
read the rbf classifier from the given ioHandler
handler | the ioHandler to be used | |
complete | if true (the default) the enclosing begin/end will be also written, otherwise only the data block will be written. |
Reimplemented from lti::classifier.
void lti::rbf::setSigmas | ( | const dvector & | s | ) |
Set the sigmas for layer 1.
object probabilities are not used
Get the test-data for training statistics
Gets the disjunct training data for object probabilities.
Training.
Mainly organization of data and training algorithms
Implements lti::supervisedInstanceClassifier.
virtual bool lti::rbf::write | ( | ioHandler & | handler, | |
const bool | complete = true | |||
) | const [virtual] |
write the rbf classifier in the given ioHandler
handler | the ioHandler to be used | |
complete | if true (the default) the enclosing begin/end will be also written, otherwise only the data block will be written. |
Reimplemented from lti::classifier.
stdLayer::distancePropFunctor* lti::rbf::distProp [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
stdLayer::dotPropFunctor* lti::rbf::dotProp [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
dmatrix lti::rbf::features [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
stdLayer::gaussActFunctor* lti::rbf::gaussAct [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
ivector lti::rbf::internToReal [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
layer1 lti::rbf::rbf1 [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
layer2 lti::rbf::rbf2 [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
std::map<int,int> lti::rbf::realToIntern [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
dvector lti::rbf::sigmas [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
stdLayer::sigmoidActFunctor* lti::rbf::sigmoidAct [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
int lti::rbf::sizeIn [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
int lti::rbf::sizeOut [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
Referenced by getDimOutputLayer().
std::map<int,dvector> lti::rbf::targetVec [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
dmatrix lti::rbf::testFeatures [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
ivector lti::rbf::testIDs [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
dmatrix lti::rbf::train2Features [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
ivector lti::rbf::train2Ids [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
ivector lti::rbf::trainID [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters
ivector lti::rbf::trainStatIDs [protected] |
Load information about the network from a '.rbf' file.
Save information about the network in a '.rbf' file
Load information about the network from a binary '.rbf' file
Save information about the network in a binary '.rbf' file
returns used parameters