Thesis docs
|
Structure containing all parameters needed for the NGGP (Normalized Generalized Gamma Process) and DP (Dirichlet Process) model. More...
#include <Params.hpp>
Public Member Functions | |
Params (double delta1=0.5, double alpha=2, double beta=2, double delta2=2, double gamma=2, double zeta=2, int BI=1000, int NI=10000, double a=1.0, double sigma=1.0, double tau=1.0, double coefficient=1, Eigen::MatrixXi W=Eigen::MatrixXi()) | |
Constructor with default parameter values. | |
Public Attributes | |
double | delta1 |
Shape parameter for the first gamma distribution prior. | |
double | alpha |
Shape parameter for the lambda_k gamma distribution. | |
double | beta |
Rate parameter for the lambda_k gamma distribution. | |
double | delta2 |
Shape parameter for the second gamma distribution prior. | |
double | gamma |
Shape parameter for the theta_kt gamma distribution. | |
double | zeta |
Rate parameter for the theta_kt gamma distribution. | |
int | BI |
Number of burn-in iterations to discard for chain convergence. | |
int | NI |
Number of iterations after burn-in for posterior sampling. | |
double | a |
Total mass parameter of the NGGP (controls number of clusters) | |
double | sigma |
Second parameter of the NGGP (controls cluster sizes) | |
double | tau |
Third parameter of the NGGP (controls tail behavior) | |
double | coefficient |
Coefficient controlling the strength of spatial dependency. | |
Eigen::MatrixXi | W |
Adjacency matrix defining spatial neighborhood structure. | |
Structure containing all parameters needed for the NGGP (Normalized Generalized Gamma Process) and DP (Dirichlet Process) model.
This structure holds all the hyperparameters, model parameters, and configuration settings required for running the DP and NGGP Bayesian nonparametric model with and without spatial dependency.
The parameters are organized into several categories:
|
inline |
Constructor with default parameter values.
delta1 | Shape parameter for first gamma prior (default: 0.5) |
alpha | Shape parameter for lambda_k gamma distribution (default: 2) |
beta | Rate parameter for lambda_k gamma distribution (default: 2) |
delta2 | Shape parameter for second gamma prior (default: 2) |
gamma | Shape parameter for theta_kt gamma distribution (default: 2) |
zeta | Rate parameter for theta_kt gamma distribution (default: 2) |
BI | Number of burn-in iterations (default: 1000) |
NI | Number of post burn-in iterations (default: 10000) |
a | NGGP total mass parameter (default: 1.0) |
sigma | NGGP second parameter (default: 1.0) |
tau | NGGP third parameter (default: 1.0) |
coefficient | Spatial dependency coefficient (default: 1) |
W | Adjacency matrix (default: empty matrix) |
double Params::a |
Total mass parameter of the NGGP (controls number of clusters)
double Params::alpha |
Shape parameter for the lambda_k gamma distribution.
double Params::beta |
Rate parameter for the lambda_k gamma distribution.
int Params::BI |
Number of burn-in iterations to discard for chain convergence.
double Params::coefficient |
Coefficient controlling the strength of spatial dependency.
double Params::delta1 |
Shape parameter for the first gamma distribution prior.
double Params::delta2 |
Shape parameter for the second gamma distribution prior.
double Params::gamma |
Shape parameter for the theta_kt gamma distribution.
int Params::NI |
Number of iterations after burn-in for posterior sampling.
double Params::sigma |
Second parameter of the NGGP (controls cluster sizes)
double Params::tau |
Third parameter of the NGGP (controls tail behavior)
Eigen::MatrixXi Params::W |
Adjacency matrix defining spatial neighborhood structure.
double Params::zeta |
Rate parameter for the theta_kt gamma distribution.