bdm::PdfDS Class Reference
Simulate data from conditional density. More...
#include <datasources.h>
Detailed Description
Simulate data from conditional density.
Data are simulated by sampling from conditional pdf.
The condition can be of two types:
- delayed output values: these are cached internally, initiated via
init_rvandinit_values. - input values: remaining RVs in the rvc of the given pdf are considered to be inputs and are declared as
urv
Note: a classical state-space model can be simulated as an instance of this case since it can be seen as a chain rule on state-evolution pdf and observation pdf. See class bdm::mprod and other offsprings of pdf.
Member Function Documentation
| void bdm::PdfDS::from_setting | ( | const Setting & | set | ) | [inline] |
Create object from the following structure
class = 'PdfDS';
pdf = configuration of bdm::pdf; % any offspring of pdf, bdm::pdf::from_setting
--- optional fields ---
init_rv = RV({'names',...},[sizes,...],[times,...]); % define what rv to initialize - typically delayed values, time=-1, etc.!
init_values = [...]; % vector of initial values corresponding to init_rv
Class does not call from bdm::DS::from_setting, names of data are taken from pdf.rv and pdf.rvc.
If init_rv is not given, init_values are set to zero.
References bdm::UI::get(), init_rv, init_values, and ipdf.
The documentation for this class was generated from the following files:
- datasources.h
- datasources.cpp
Generated on Fri Aug 27 16:54:40 2010 for mixpp by
1.6.0
