LEARN_MODEL_LW_PI_WS Struct Reference

This structure describes workspace for directly learning the policy model parameters using locally weighted linear regression. More...

#include <ccl_learn_policy.h>

Data Fields
gsl_vector *	g
	Col vector of YPhit.
gsl_matrix *	Y_N
	Normalised Y.
gsl_vector *	YN_vec
	Vector view of Y_N.
gsl_matrix *	Y_Phit
	Dot product of Y * WPhi'.
gsl_matrix *	ones
	Matrix of all ones.
gsl_matrix *	Y_
	Output variable.
gsl_matrix *	H
	Accumulated Hessian.
gsl_matrix *	Phi
	Feature matrix.
gsl_vector *	Phi_vec
	Feature vector.
gsl_matrix *	Phi_vec_T
	Transpose of Feature vector.
gsl_matrix *	YN_Phit
	Dot prodcut of YN * Phit.
gsl_vector *	YN_Phi_vec
	Vector view of YN_Phit.
gsl_matrix *	YN_Phi_vec_T
	Transpose of YN_Phi_vec.
gsl_matrix *	vv
	Dot product of v * v.
gsl_matrix *	WX_
	Dot product of W * X_.
gsl_vector *	WX_row
	Row vector of WX_.
gsl_matrix *	WPhi
	Dot product of W * Phi.
gsl_matrix *	WPhi_T
	Transpose of WPhi.
gsl_matrix *	pinvH1
	Peudo inverse of H1.
gsl_vector *	V
	Eigen vector of H.
gsl_vector *	r
	Normalisation scaler for YN.
gsl_matrix *	r_rep
	Replication matrix of r.
gsl_matrix *	D
	Eigen values of H.
int *	idx
	Index.
gsl_vector *	w_vec
	Vector view of model parameter w for each center.
gsl_matrix *	w_
	Model parameter w for each center.
gsl_matrix *	w_T
	Transpose of model parameter w for each center.
gsl_matrix *	w [NUM_CENTRES]
	Model parameter w for all centers.

Detailed Description

This structure describes workspace for directly learning the policy model parameters using locally weighted linear regression.

This structure defines the worspace variables and initialises with their dimensionalities. It should always initialised with ccl_learn_model_lw_pi_ws_alloc, and destroyed with ccl_learn_model_lw_pi_ws_free.

---

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

• include/ccl_learn_policy.h