\(\newcommand{\B}[1]{ {\bf #1} }\) \(\newcommand{\R}[1]{ {\rm #1} }\) \(\newcommand{\W}[1]{ \; #1 \; }\)
var_table#
View page sourceIdentifying Model Variables#
Discussion#
The variable
table maps the var_id
to information that specifies exactly which
model_variable it corresponds to.
You can use this information to interpret a fit_var_table
created by the fit_command ,
or to create a truth_var_table
for use as input to the simulate_command .
var_id#
This column has type integer
and is the primary key for this table.
Its initial value is zero, and it increments by one for each row.
var_type#
The column has type text
and has one of the following possible values:
mulstd_value#
This variable multiplies the standard deviation in the mulstd_value_prior for the corresponding smooth_id ; see smoothing standard deviation multiplier .
mulstd_dage#
This variable multiplies the standard deviation in the mulstd_dage_prior for the corresponding smooth_id ; see smoothing standard deviation multiplier .
mulstd_dtime#
This variable multiplies the standard deviation in the mulstd_dtime_prior for the corresponding smooth_id ; see smoothing standard deviation multiplier .
rate#
This variable is rate for the corresponding age_id , time_id , node_id , and rate_id . If the node_id is the parent_node_id , it is a fixed effect . Otherwise, it is a random effect .
mulcov_rate_value#
This variable corresponds to a group or subgroup covariate multiplier that affects the rate for the corresponding age_id , time_id , rate_id , and covariate_id .
mulcov_meas_value#
This variable corresponds to a group or subgroup covariate multiplier that affects the measurement value for the corresponding age_id , time_id , integrand_id , and covariate_id .
mulcov_meas_noise#
This variable corresponds to a group covariate multiplier that affects the measurement value age_id , time_id , integrand_id , and covariate_id . There are no subgroup covariate multipliers of this type.
smooth_id#
If var_type is
mulstd_value
, mulstd_dage
or mulstd_dtime
,
this is the smooth_id for the
smoothing that this variable multiplies.
Otherwise, the specified smoothing is used for the priors
for this variable.
To be specific,
the corresponding priors can be found by matching
smooth_id , age_id , and time_id in the
smooth_grid_table .
age_id#
If var_type is
mulstd_value
, mulstd_dage
, mulstd_dtime
,
this column is null
.
Otherwise this is the age_id that identifies
the age for the corresponding variable.
time_id#
If var_type is
mulstd_value
, mulstd_dage
, mulstd_dtime
,
this column is null
.
Otherwise this is the time_id that identifies
the time for the corresponding variable.
Otherwise, this column is null
.
node_id#
If var_type is rate
,
this is the node_id that the rate is for.
Otherwise, this column is null
.
rate_id#
If var_type is rate
or mulcov_rate_value
,
this is the rate_id for the specific rate.
Otherwise, this column is null
.
integrand_id#
If var_type is
mulcov_meas_value
or mulcov_meas_noise
,
this is the integrand_id
that this variable applies to.
Otherwise, this column is null
.
covariate_id#
If var_type is
mulcov_rate_value
, mulcov_meas_value
or mulcov_meas_noise
,
this is the covariate_id
that this variable applies to.
Otherwise, this column is null
.
mulcov_id#
If var_type is
mulcov_rate_value
, mulcov_meas_value
or mulcov_meas_noise
,
this is the mulcov_id
that this variable applies to.
Otherwise, this column is null
.
group_id#
mulcov_meas_noise#
If var_type is mulcov_meas_noise
,
group_id is the mulcov table
group_id
that this fixed effect variable applies to.
mulcov_rate_value, mulcov_meas_value#
If var_type is mulcov_rate_value
or mulcov_meas_value
.
In this case either group_id or subgroup_id
(but not both) is not null.
If group_id is not null,
it is the mulcov table
group_id
that this fixed effect applies to.
Otherwise#
group_id must be null.
subgroup_id#
mulcov_rate_value, mulcov_meas_value#
If var_type is mulcov_rate_value
or mulcov_meas_value
.
In this case either group_id or subgroup_id
(but not both) is not null.
If subgroup_id is not null,
it is the subgroup_id
that this random effect applies to.
Otherwise#
subgroup_id must be null.
Example#
See the init_command.py example and test.