Description Usage Arguments Details Value Examples
This function creates a list with all required objects at the global level needed by the sema algorithm
1 2 | create_theta_main(n_fixed, n_random, start_resid_var = 1,
start_random_var = 1, start_fixed_coef = NULL, prior_n, prior_j)
|
n_fixed |
Number of fixed effects. |
n_random |
Number of random effects. |
start_resid_var |
Start values residual variance, default start value equals 1. |
start_random_var |
Start values variance of random effects,
default start values equal 1. Make sure that the length of
|
start_fixed_coef |
The default is set to NULL, when no start values
are provided, |
prior_n |
A scalar indicating the weight of the start value of the residual variance. |
prior_j |
A scalar indicating the weight of the start values of the fixed effects coefficients and random effects variance. |
The function creates a list with objects used by the sema algorithm, most are not used for interpretation. The user can provide start values to enhance the algorithm's performance. In that case, the function also requires the weight of those start values, when no starting values are given default values are used.
A list which contains: the fixed effects coefficients,
fixed_coef_hat
; the Complete Data Sufficient Statistic of the
fixed effects, t1
; the variance of the random effects
random_var_hat
; the Complete Data Sufficent Statistics of the
variance of random effects, t2
; the residual variance,
resid_var_hat
; the Complete Data Sufficiant Statistic of
the residual variance, t3
, the number of observations, n
;
the number of individuals, j
; the average dependent variable,
which can be used for prediction purposes but it is not needed for the
algorithm, y
; the square matrix of squared fixed effects, which is
only updated as long as the matrix is not yet invertible, once inverted,
only the inverted matrix is updated, x_sq
; the product of fixed
effects and dependent variable, xy_vector
; the inverse of the
x_sq
matrix, x_inv
.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 | ## Create a list of objects required to fit the multilevel model using
## sema:
## NOTE: default start values fixed effect coefficients, residual variance
## and random effects variance is equal to 1. When this function is used
## outside the sema_fit functions, prior_n and prior_j, i.e., weight given
## to the start values should be entered. Within the sema_fit functions
## defaults are given.
model_statistics <- create_theta_main(n_fixed = 5,
n_random = 3,
start_resid_var = 1,
start_random_var = 1,
start_fixed_coef = NULL,
prior_n = 0,
prior_j = 0)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.