BALL  1.4.2
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Properties Friends Macros Groups Pages
List of all members
BALL::QSAR::ALLModel Class Reference

#include <BALL/QSAR/allModel.h>

Inheritance diagram for BALL::QSAR::ALLModel:
BALL::QSAR::NonLinearModel BALL::QSAR::RegressionModel BALL::QSAR::Model BALL::QSAR::KNNModel

Public Member Functions

Constructors and Destructors
 ALLModel (const QSARData &q, double kw=4)
 
 ~ALLModel ()
 
- Public Member Functions inherited from BALL::QSAR::NonLinearModel
 NonLinearModel (const QSARData &q)
 
 ~NonLinearModel ()
 

Protected Attributes

Attributes
double kw_
 
double lambda_
 

Accessors

void setKw (double kw)
 
virtual Vector< doublepredict (const vector< double > &substance, bool transform=1)
 
void train ()
 
virtual bool optimizeParameters (int d, int no_steps)
 
double getKw ()
 
virtual void setParameters (vector< double > &v)
 
virtual vector< doublegetParameters () const
 
void saveToFile (string filename)
 
void readFromFile (string filename)
 
virtual void calculateWeights (BALL::Matrix< double > &dist, BALL::Vector< double > &w)
 
void calculateXX (BALL::Vector< double > &w, BALL::Matrix< double > &res)
 
void calculateXY (BALL::Vector< double > &w, BALL::Matrix< double > &res)
 
void calculateEuclDistanceMatrix (BALL::Matrix< double > &m1, BALL::Matrix< double > &m2, BALL::Matrix< double > &output)
 

Additional Inherited Members

- Public Attributes inherited from BALL::QSAR::RegressionModel
RegressionValidationvalidation
 
- Public Attributes inherited from BALL::QSAR::Model
const QSARDatadata
 
Validationmodel_val
 
- Protected Member Functions inherited from BALL::QSAR::NonLinearModel
virtual void calculateOffsets ()
 
- Protected Member Functions inherited from BALL::QSAR::RegressionModel
void readDescriptorInformationFromFile (std::ifstream &in, int no_descriptors, bool transformation, int no_coefficients)
 
void saveDescriptorInformationToFile (std::ofstream &out)
 

Detailed Description

class for automated lazy learning (ALL-QSAR)

Definition at line 40 of file allModel.h.

Constructor & Destructor Documentation

BALL::QSAR::ALLModel::ALLModel ( const QSARData q,
double  kw = 4 
)
BALL::QSAR::ALLModel::~ALLModel ( )

Member Function Documentation

void BALL::QSAR::ALLModel::calculateEuclDistanceMatrix ( BALL::Matrix< double > &  m1,
BALL::Matrix< double > &  m2,
BALL::Matrix< double > &  output 
)
protected

calculates pairwise euclidean distance between all substances of m1 and m2 and saves them to BALL::Matrix<double> output

virtual void BALL::QSAR::ALLModel::calculateWeights ( BALL::Matrix< double > &  dist,
BALL::Vector< double > &  w 
)
protectedvirtual

Reimplemented in BALL::QSAR::KNNModel.

void BALL::QSAR::ALLModel::calculateXX ( BALL::Vector< double > &  w,
BALL::Matrix< double > &  res 
)
protected

calculates weighted BALL::Matrix<double> X^T*X

void BALL::QSAR::ALLModel::calculateXY ( BALL::Vector< double > &  w,
BALL::Matrix< double > &  res 
)
protected

calculates weighted BALL::Matrix<double> X^T*Y

double BALL::QSAR::ALLModel::getKw ( )

returns the current kernel width

virtual vector<double> BALL::QSAR::ALLModel::getParameters ( ) const
virtual

Reimplemented from BALL::QSAR::Model.

Reimplemented in BALL::QSAR::KNNModel.

virtual bool BALL::QSAR::ALLModel::optimizeParameters ( int  d,
int  no_steps 
)
virtual

Tries to find the optimal kernel width for the current data. Therefore some training data must have been read by the connected QSARData object before running this method.

Reimplemented from BALL::QSAR::Model.

Reimplemented in BALL::QSAR::KNNModel.

virtual Vector<double> BALL::QSAR::ALLModel::predict ( const vector< double > &  substance,
bool  transform = 1 
)
virtual
Predicts the activities of a given substance
Parameters
substancethe substance which activity is to be predicted in form of a vecor containing the values for all descriptors (if neccessary, relevant descriptors will be selected automatically)
transformdetermines whether the values for each descriptor of the given substance should be transformed before prediction of activity.
If (transform==1): each descriptor value is transformed according to the centering of the respective column of QSARData.descriptor_matrix used to train this model.
If the substance to be predicted is part of the same input data (e.g. same SD-file) as the training data (as is the case during cross validation), transform should therefore be set to 0.
Returns
a RowVector containing one value for each predicted activity

Implements BALL::QSAR::Model.

void BALL::QSAR::ALLModel::readFromFile ( string  filename)
virtual

reconstruct a saved Model from a file

Reimplemented from BALL::QSAR::RegressionModel.

void BALL::QSAR::ALLModel::saveToFile ( string  filename)
virtual

save Model to a file

Reimplemented from BALL::QSAR::RegressionModel.

void BALL::QSAR::ALLModel::setKw ( double  kw)
virtual void BALL::QSAR::ALLModel::setParameters ( vector< double > &  )
virtual

sets the model parameters according to the given values.

Reimplemented from BALL::QSAR::Model.

Reimplemented in BALL::QSAR::KNNModel.

void BALL::QSAR::ALLModel::train ( )
inlinevirtual

automated lazy learning does not have a seperate training step

Implements BALL::QSAR::Model.

Definition at line 59 of file allModel.h.

Member Data Documentation

double BALL::QSAR::ALLModel::kw_
protected

kernel width

Definition at line 99 of file allModel.h.

double BALL::QSAR::ALLModel::lambda_
protected

Definition at line 100 of file allModel.h.