Regression Model
Use the drop-down menu to select a Bayesian method for the estimation of marker variables effects.
Available options are listed in the following table:
|
Option |
Description |
|
BayesA |
Regression coefficients are assigned the Scaled-t density prior, which has higher mass at zero and thicker tails than the normal (Gaussian) density. The shrinkage of coefficient estimates is dependent on their effect-size. |
|
BayesB |
Regression coefficients are assigned a mixture of two finite mixture priors, a mixture of a point mass at zero and a Scaled-t density otherwise. |
|
BayesC |
Regression coefficients are assigned a mixture of two finite mixture priors, a mixture of a point mass at zero and a Gaussian density otherwise. |
|
BayesianLASSO |
Regression coefficients are assigned the Double Exponential (DE) density prior, which has higher mass at zero and thicker tails than the normal (Gaussian) density. The shrinkage of coefficient estimates is dependent on their effect-size. |
|
BRR |
Regression coefficients are assigned the Gaussian density prior, which induces shrinkage of coefficient estimates like in Ridge Regression. |
To Specify the Regression Model:
| 8 | Select the desired measure using the drop-down menu. |