bdm::BMEF Class Reference
Estimator for Exponential family. More...
#include <exp_family.h>
Inheritance diagram for bdm::BMEF:

Public Member Functions | |
| BMEF (double frg0=1.0) | |
| Default constructor (=empty constructor). | |
| BMEF (const BMEF &B) | |
| Copy constructor. | |
| virtual void | set_statistics (const BMEF *BM0) NOT_IMPLEMENTED_VOID |
| get statistics from another model | |
| virtual void | bayes_weighted (const vec &data, const vec &cond=empty_vec, const double w=1.0) |
| Weighted update of sufficient statistics (Bayes rule). | |
| void | bayes (const vec &yt, const vec &cond=empty_vec) |
| virtual void | flatten (const BMEF *B, double weight=1.0) NOT_IMPLEMENTED_VOID |
| Flatten the posterior according to the given BMEF (of the same type!). | |
| void | to_setting (Setting &set) const |
| void | from_setting (const Setting &set) |
| void | validate () |
Public Attributes | |
| double | frg |
| forgetting factor | |
Protected Attributes | |
| double | last_lognc |
cached value of lognc() in the previous step (used in evaluation of ll ) | |
| double | frg_sched_factor |
factor k = [0..1] for scheduling of forgetting factor: , default 0 | |
Detailed Description
Estimator for Exponential family.
Member Function Documentation
| void bdm::BMEF::from_setting | ( | const Setting & | set | ) | [inline, virtual] |
Create object from the following structure
class = 'BMEF'; --- optional fields --- frg = []; % forgetting factor frg_sched_factor = []; % factor for scheduling of forgetting factor: a number from [0..1] --- inherited fields --- bdm::BM::from_setting
frg = 1; % default forgetting factor
frg_sched_factor = 0;
Reimplemented from bdm::BM.
Reimplemented in bdm::ARX, bdm::ARXg, bdm::MixEF, and bdm::multiBM.
The documentation for this class was generated from the following files:
- exp_family.h
- exp_family.cpp
Generated on 2 Dec 2013 for mixpp by
1.4.7

, default 0