init_lm_hs {countSTAR} | R Documentation |
Initialize linear regression parameters assuming a horseshoe prior
Description
Initialize the parameters for a linear regression model assuming a
horseshoe prior for the (non-intercept) coefficients. The number of predictors
p
may exceed the number of observations n
.
Usage
init_lm_hs(y, X, X_test = NULL)
Arguments
y |
|
X |
|
X_test |
|
Value
a named list params
containing at least
-
mu
: vector of conditional means (fitted values) -
sigma
: the conditional standard deviation -
coefficients
: a named list of parameters that determinemu
Additionally, if X_test is not NULL, then the list includes an element
mu_test
, the vector of conditional means at the test points
Note
The parameters in coefficients
are:
-
beta
: thep x 1
vector of regression coefficients -
sigma_beta
: thep x 1
vector of regression coefficient standard deviations (local scale parameters) -
xi_sigma_beta
: thep x 1
vector of parameter-expansion variables forsigma_beta
-
lambda_beta
: the global scale parameter -
xi_lambda_beta
: the parameter-expansion variable forlambda_beta
components ofbeta