| Type: | Package |
| Title: | Bayesian Grouped Weighted Quantile Sum Regression |
| Version: | 0.1.1 |
| Author: | David Wheeler, Matthew Carli |
| Maintainer: | Matthew Carli <carlimm@mymail.vcu.edu> |
| Description: | Fits Bayesian grouped weighted quantile sum (BGWQS) regressions for one or more chemical groups with binary outcomes. Wheeler DC et al. (2019) <doi:10.1016/j.sste.2019.100286>. |
| License: | GPL-3 |
| Encoding: | UTF-8 |
| LazyData: | true |
| RoxygenNote: | 7.1.1 |
| Depends: | R (≥ 4.0.0) |
| SystemRequirements: | JAGS |
| Imports: | coda, stats, rjags, stringr, plyr |
| Suggests: | testthat |
| NeedsCompilation: | no |
| Packaged: | 2022-01-20 22:32:42 UTC; Matt Carli |
| Repository: | CRAN |
| Date/Publication: | 2022-01-20 22:42:54 UTC |
Bayesian Grouped WQS Regression
Description
This function fits a Bayesian grouped weighted quantile sum (BGWQS) regression model.
Usage
bgwqs.fit(
y,
x,
z,
x.s,
n.quantiles = 4,
working.dir,
n.chains = 1,
n.iter = 10000,
n.burnin = 5000,
n.thin = 1,
n.adapt = 500,
DIC = FALSE
)
Arguments
y |
A vector containing outcomes. |
x |
A matrix of component data. |
z |
A vector or matrix of controlling covariates. |
x.s |
A vector of the number of components in each index. |
n.quantiles |
The number of quantiles to apply to the component data. |
working.dir |
A file path to the directory. |
n.chains |
The number of Markov chains; must be a positive integer. |
n.iter |
The number of total iterations per chain, including burn in. |
n.burnin |
The number of iterations to discard at the beginning. |
n.thin |
The thinning rate; must be a positive integer. |
n.adapt |
The number of adaption iterations. |
DIC |
Logical; whether or not the user desires the function to return DIC. |
Value
A list which includes BUGS output, sample chains post-burnin, and convergence test results.
Examples
## Not run:
data("simdata")
group_list <- list(c("pcb_118", "pcb_138", "pcb_153", "pcb_180", "pcb_192"),
c("as", "cu", "pb", "sn"),
c("carbaryl", "propoxur", "methoxychlor", "diazinon", "chlorpyrifos"))
x.s <- make.x.s(simdata, 3, group_list)
X <- make.X(simdata, 3, group_list)
Y <- simdata$Y
work_dir <- tempdir()
results <- bgwqs.fit(y = Y, x = X, x.s = x.s, n.quantiles=4,
working.dir = work_dir,
n.chains = 1, n.iter = 10000, n.burnin = 5000, n.thin = 1, n.adapt = 500)
## End(Not run)
Forms matrix of components
Description
This function returns a matrix of component variables, X. The user can specify the desired chemicals and order by creating a list of string vectors, each vector containing the variable names of all desired elements of that group.
Usage
make.X(df, num.groups, groups)
Arguments
df |
A dataframe containing named component variables |
num.groups |
An integer representing the number of component groups desired |
groups |
A list, each item in the list being a string vector of variable names for one component group |
Value
A matrix of component variables
Examples
data("simdata")
group_list <- list(c("pcb_118", "pcb_138", "pcb_153", "pcb_180", "pcb_192"),
c("as", "cu", "pb", "sn"),
c("carbaryl", "propoxur", "methoxychlor", "diazinon", "chlorpyrifos"))
X <- make.X(simdata, 3, group_list)
X
Forms component group ID vector of X
Description
This function returns a vector which lets WQS.fit know the size and order of groups in X
Usage
make.x.s(df, num.groups, groups)
Arguments
df |
A dataframe containing named component variables |
num.groups |
An integer representing the number of component groups desired |
groups |
A list, each item in the list being a string vector of variable names for one component group |
Value
A vector of integers, each integer relating how many columns are in each group
Examples
data("simdata")
group_list <- list(c("pcb_118", "pcb_138", "pcb_153", "pcb_180", "pcb_192"),
c("as", "cu", "pb", "sn"),
c("carbaryl", "propoxur", "methoxychlor", "diazinon", "chlorpyrifos"))
x.s <- make.x.s(simdata, 3, group_list)
x.s
Simulated data of chemical concentrations and one binary outcome variable
Description
Data were simulated to have 0.7 in-group correlation and 0.3 between-group correlation. There are three groups, with the third being significantly correlated to the outcome variable.
Usage
simdata
Format
A data frame with 1000 rows and 15 variables:
- pcb_118
a numeric vector; part of group 1
- pcb_138
a numeric vector; part of group 1
- pcb_153
a numeric vector; part of group 1
- pcb_180
a numeric vector; part of group 1
- pcb_192
a numeric vector; part of group 1
- as
a numeric vector; part of group 2
- cu
a numeric vector; part of group 2
- pb
a numeric vector; part of group 2
- sn
a numeric vector; part of group 2
- carbaryl
a numeric vector; part of group 3
- propoxur
a numeric vector; part of group 3
- methoxychlor
a numeric vector; part of group 3
- diazinon
a numeric vector; part of group 3
- chlorpyrifos
a numeric vector; part of group 3
- Y
a numeric vector; the outcome variable
Generates Plots of weights by group
Description
This function takes the object created by the bgwqs.fit function and a vector of group names and generates a random forest variable importance plot for each group. The weights in each group are listed in descending order.
Usage
weight.plot(fit.object, group.names, group.list, x.s)
Arguments
fit.object |
The object that is returned by the bgwqs.fit function |
group.names |
A string vector containing the name of each group included in the BGWQS regression. Will be used for plot titles. |
group.list |
A list, each item in the list being a string vector of variable names for one component group. |
x.s |
A vector of the number of components in each index. |
Value
A plot for each group of the BGWQS regression