BALL::QSAR::LatentVariableModel Class Reference

#include <BALL/QSAR/latentVariableModel.h>

Inheritance diagram for BALL::QSAR::LatentVariableModel:

[legend]

List of all members.

Public Member Functions
	LatentVariableModel ()
const Matrix< double > *	getLatentVariables ()
const Matrix< double > *	getWeights ()
const Matrix< double > *	getLoadings ()
Protected Attributes
Matrix< double >	latent_variables_
Matrix< double >	loadings_
Matrix< double >	weights_
Friends
class	RegressionValidation

---

Detailed Description

Interface-class for all latente variable models (PLS, PCR, ...). It is used to store and access the latente variables (PLS/PCR components), the loading matrix and the weight matrix in the same way for all such models.

Definition at line 40 of file latentVariableModel.h.

---

Constructor & Destructor Documentation

BALL::QSAR::LatentVariableModel::LatentVariableModel

(

)

---

Member Function Documentation

const Matrix<double>* BALL::QSAR::LatentVariableModel::getLatentVariables

(

)

Return the latent variable matrix as computed by the last training.
It contains one loading vector in each column. Each loading vector contains the importance of each feature for a latente variable.

const Matrix<double>* BALL::QSAR::LatentVariableModel::getLoadings

(

)

Return the loading matrix as computed by the last training.
It contains one loading vector in each column. Each loading vector contains the importance of each feature for a latente variable.

const Matrix<double>* BALL::QSAR::LatentVariableModel::getWeights

(

)

Return the weights matrix as computed by the last training.
It contains one weight vector in each column. Each weight vector contains the importance of each latente variable for modelling the result.

---

Friends And Related Function Documentation

friend class RegressionValidation [friend]

Definition at line 68 of file latentVariableModel.h.

---

Member Data Documentation

Matrix<double> BALL::QSAR::LatentVariableModel::latent_variables_ [protected]

The loading-matrix which contains one loading vector in each column. Each loading vector contains the importance of each feature for a latente variable.

Definition at line 60 of file latentVariableModel.h.

Matrix<double> BALL::QSAR::LatentVariableModel::loadings_ [protected]

The loading-matrix which contains one loading vector in each column. Each loading vector contains the importance of each feature for a latente variable.

Definition at line 63 of file latentVariableModel.h.

Matrix<double> BALL::QSAR::LatentVariableModel::weights_ [protected]

The weights-matrix which contains one weight vector in each column. Each weight vector contains the importance of each latente variable for modelling the result.

Definition at line 66 of file latentVariableModel.h.