Skip to contents

Create Parameters for Simulation with Uncertainty

Source: R/simpar.R
simpar.Rd

Create Parameters for Simulation with Uncertainty

Usage

simpar(
  nsim,
  theta,
  covar,
  omega,
  sigma,
  odf = NULL,
  sdf = NULL,
  digits = 4,
  min = -Inf,
  max = Inf
)

Arguments

nsim

scalar numeric specifying the number of sets to attempt

theta

vector of point estimates of fixed effect parameters

covar

variance-covariance matrix for fixed effect parameters

omega

list of variance-covariance matrices for first level random effects

sigma

list of variance-covariance matrices for second level random effects

odf

vector of omega degrees of freedom, one per matrix

sdf

vector of sigma degrees of freedom, one per matrix

digits

number of significant digits to include in output

min

lower limit for parameter estimates

max

upper limit for parameter estimates

Value

matrix, with column names indicating parameters, and row names indicating set number before filtering by min and max.

Details

If min or max are non-default (see below), you may want to set nsim marginally higher to allow for dropped sets. covar is coerced to matrix using as.matrix.

If omega and sigma are not lists, they are coerced using list. Then each element is coerced using as.matrix.

By default, each element in odf and sdf will be the length (number of elements) in the corresponding matrix.

min and max may be given as scalar values, in which case they apply to all parameters (as do the defaults). Alternatively, the first n limits may be specified as a vector, in which case the remaining (if any) will be the default. If any simulated parameter does not fall between its limits, inclusive, the entire parameter set (row) is dropped from the result, with warning.

Examples


set.seed(100)
simpar(
 nsim=10,
 theta=c(13,75,1),
 covar=matrix(c(10,7,2,7,30,1,2,1,0.5),ncol=3,nrow=3),
 omega=list(
   0.1,
   matrix(c(0.04,0.02,0.02,0.04),ncol=2,nrow=2)
 ),
 odf=c(50,20),
 sigma=list(0.04,1),
 sdf=c(99,99),
 min=rep(0,3),
 max=rep(90,3)
)
#>     TH.1  TH.2   TH.3   OM1.1   OM2.2     OM3.2   OM3.3   SG1.1  SG2.2
#> 1  14.13 77.63 1.0640 0.10390 0.04442  0.007829 0.03910 0.03569 1.0690
#> 2  12.98 74.21 1.2430 0.09322 0.05781  0.048300 0.07054 0.04070 1.0910
#> 3  12.59 75.60 0.9816 0.11000 0.06001  0.009597 0.03225 0.04675 1.0330
#> 4  13.38 69.56 1.4090 0.11790 0.10790  0.061510 0.06000 0.03660 0.9531
#> 5  13.18 74.26 0.8007 0.11190 0.08363  0.046380 0.07052 0.04134 1.2040
#> 6  12.40 73.30 0.7659 0.09395 0.03944  0.017340 0.07900 0.03551 0.9038
#> 7  13.02 78.48 0.7257 0.08317 0.05386  0.027630 0.04757 0.04314 0.8023
#> 8  13.11 70.69 1.1650 0.08464 0.04884  0.041840 0.08535 0.04101 0.8885
#> 9  12.07 80.26 0.3608 0.10420 0.03331 -0.006786 0.06227 0.03728 0.8611
#> 10 19.73 74.93 2.5130 0.06927 0.06935  0.050570 0.05766 0.03791 0.9706
simpar(
 nsim=1,
 theta=c(13,75,1),
 covar=matrix(c(10,7,2,7,30,1,2,1,0.5),ncol=3,nrow=3),
 omega=list(
   0.1,
   matrix(c(0.04,0.02,0.02,0.04),ncol=2,nrow=2)
 ),
 odf=c(50,20),
 sigma=list(0.04,1),
 sdf=c(99,99),
 min=rep(0,3),
 max=rep(90,3)
)
#>    TH.1  TH.2 TH.3   OM1.1   OM2.2   OM3.2   OM3.3  SG1.1 SG2.2
#> 1 15.18 75.47 1.29 0.09543 0.03465 0.01318 0.07248 0.0442  1.26
simpar(
 nsim=1,
 theta=c(13,75,1),
 covar=matrix(c(10,7,2,7,30,1,2,1,0.5),ncol=3,nrow=3),
 omega=list(
   0.1,
   matrix(c(0.04,0.02,0.02,0.04),ncol=2,nrow=2)
 ),
 odf=c(50,20),
 sigma=list(0.04,1),
 sdf=c(99,99),
 min=Inf,
 max=-1
)
#> Warning: 1 of 1 rows dropped
#>      TH.1 TH.2 TH.3 OM1.1 OM2.2 OM3.2 OM3.3 SG1.1 SG2.2