Description Usage Arguments Value See Also Examples
From a matrix of bootstrapped coefficients, calculate and plot odds ratios, 95 and Wald p-values for all coefficients, then plot.
1 2 3 | multi.plot.ors(coef.list, label.data = NULL, remove.vars = NULL,
round.vars = NULL, round.digits = NULL, out.strings.list,
delete.row = "none", yval.offset = 0.25)
|
coef.list |
List of matrices including coefficients from bootstrapped models (columns = coefficients). |
label.data |
If desired, data frame with two columns, variable and var.label, containing variable names and strings to use in plot labels, respectively. Default is NULL. |
remove.vars |
Character vector of variable names to **not** include in calculations/plots.
Defaults to NULL (show all variables). Passed to |
round.vars |
Character vector of variable names whose results should be rounded to something
other than two decimal places. Useful for variables with very small changes in odds for one-unit
change in variable. Defaults to NULL. Passed to |
round.digits |
Integer; number of digits to round [round.vars] to.
Passed to |
out.strings.list |
List of character vectors to label outcome comparisons. |
delete.row |
Row to delete from plots and calculations. Used in situations where models are run twice with different reference levels; in this case, one comparison is redundant (eg, 'B vs. A' is reciprocal of 'A vs. B'). |
yval.offset |
Numeric; amount to offset lines for each outcome level in final plot. |
List of 1) or.data
, a data frame containing odds ratios, confidence
limits, p-values and accompanying information; 2) or.plot
, a ggplot2 object
which plots ORs and CIs for all variables and outcome comparisons included, adding
p-values to axis labels.
ggplot2.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 | df <- data.frame(id = sample(1:20, size = 100, replace = TRUE),
x1 = rnorm(n = 100),
x2 = rbinom(p = 0.75, n = 100, size = 1),
y = sample(LETTERS[1:3], size = 100, replace = TRUE))
df <- df[order(df$id),]
df$time <- unlist(lapply(1:length(unique(df$id)),
FUN = function(idnum){ 1:nrow(df[df$id == unique(df$id)[idnum],]) }))
## Using create.sampdata(), generate list of cluster bootstrapped data sets
bootdata.list <- create.sampdata(org.data = df,
id.var = 'id',
n.sets = 25)
## Fit model to original and bootstrapped data frame,
## saving errors and warnings to .txt file
boot.fits.a <- multi.bootstrap(org.data = df,
data.sets = bootdata.list,
ref.outcome = grep('A', levels(df$y)),
multi.form = as.formula('y ~ x1 + x2'))
## Create matrices of coefficients for all bootstrap fits
boot.matrix.a <- do.call(rbind,
lapply(boot.fits.a$boot.models,
FUN = function(x){ x@coefficients }))
## Calculate and plot odds ratios and CIs, Wald p-values for x2
covariate.ors <- multi.plot.ors(coef.list = list(boot.matrix.a),
out.strings = list(c('B vs A', 'C vs A')),
remove.vars = 'x1')
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.