8.1 Double hierarchical models (DHGLMs)
8.1.1 DHGLM
<- simulate_population(
squid_data data_structure=make_structure("individual(50)",N=10),
parameters = list(
individual = list(
names = c("ind_int","ind_slopes","ind_lnsd"),
mean =c(0, 0, 0.5),
vcov = c(0.5, 0.1, 0.1),
beta = c(1, 0, 0),
functions=c(NA,NA,"exp")
),observation =list(
names="environment",
beta=0.2
),residual = list(
vcov = 1,
beta = 0
),interactions = list(
names = c("ind_slopes:environment","ind_lnsd:residual"),
beta = c(1, 1)
)
) )
8.1.2 Bivariate DHGLM
<- simulate_population(
squid_data n_response=2,
parameters = list(
individual = list(
names = c("ind_int1","ind_slopes1","ind_lnsd1","ind_int2","ind_slopes2","ind_lnsd2"),
mean =c(0, 0, 0.5,0,0,1),
vcov = c(0.5,0.1,0.1,0.4,0.2,0.05),
beta = matrix(c( 1,0,
0,0,
0,0,
0,1,
0,0,
0,0
byrow=TRUE,ncol=2),
),
functions=c(NA,NA,"exp",NA,NA,"exp")
),observation =list(
names="environment",
beta=matrix(c(0.2,-0.3),ncol=2)
),residual = list(
names = c("residual1","residual2"),
vcov = c(1,1),
beta = matrix(c(0,0,0,0),ncol=2)
),interactions = list(
names = c("ind_slopes1:environment","ind_lnsd1:residual1","ind_slopes2:environment","ind_lnsd2:residual2"),
beta = matrix(c(1,1,0,0,0,0,1,1),byrow=TRUE,ncol=2)
)
),data_structure=make_structure("individual(50)",N=10)
)