| Title: | Bayesian Framework for Computational Modeling |
|---|---|
| Description: | Derived from the work of Kruschke (2015, <ISBN:9780124058880>), the present package aims to provide a framework for conducting Bayesian analysis using Markov chain Monte Carlo (MCMC) sampling utilizing the Just Another Gibbs Sampler ('JAGS', Plummer, 2003, <https://mcmc-jags.sourceforge.io>). The initial version includes several modules for conducting Bayesian equivalents of chi-squared tests, analysis of variance (ANOVA), multiple (hierarchical) regression, softmax regression, and for fitting data (e.g., structural equation modeling). |
| Authors: | Øystein Olav Skaar [aut, cre] |
| Maintainer: | Øystein Olav Skaar <[email protected]> |
| License: | MIT + file LICENSE |
| Version: | 0.4.2 |
| Built: | 2024-11-17 06:52:25 UTC |
| Source: | CRAN |
Add names to columns from naming list
AddNames( par, job.names, job.group = NULL, keep.par = TRUE, names.only = FALSE, ... )AddNames( par, job.names, job.group = NULL, keep.par = TRUE, names.only = FALSE, ... )
par |
defined parameter to analyze (e.g., "cor[1,2]") |
job.names |
names of all parameters in analysis, Default: NULL |
job.group |
for some hierarchical models with several layers of parameter names (e.g., latent and observed parameters), Default: NULL |
keep.par |
logical, indicating whether or not to keep parameter name (e.g., "cor[1,2]"), Default: TRUE |
names.only |
logical, indicating whether or not to return vector (TRUE) or string with separator (e.g., "cor[1,2]: A vs. B"), Default: FALSE |
... |
further arguments passed to or from other methods |
par <- "cor[1,2]" job.names <- c("A","B") AddNames(par, job.names, keep.par = TRUE) # [1] "cor[1,2]: A vs. B" AddNames(par, job.names, keep.par = FALSE) # [1] "A vs. B" AddNames(par, job.names, names.only = TRUE) # [1] "A" "B"par <- "cor[1,2]" job.names <- c("A","B") AddNames(par, job.names, keep.par = TRUE) # [1] "cor[1,2]: A vs. B" AddNames(par, job.names, keep.par = FALSE) # [1] "A vs. B" AddNames(par, job.names, names.only = TRUE) # [1] "A" "B"
main settings for bfw
bfw( job.title = NULL, job.group = NULL, jags.model, jags.seed = NULL, jags.method = NULL, jags.chains = NULL, custom.function = NULL, custom.model = NULL, params = NULL, saved.steps = 10000, thinned.steps = 1, adapt.steps = NULL, burnin.steps = NULL, initial.list = list(), custom.name = NULL, project.name = "Project", project.dir = "Results/", project.data = NULL, time.stamp = TRUE, save.data = FALSE, data.set = "AllData", data.format = "csv", raw.data = FALSE, run.robust = FALSE, merge.MCMC = FALSE, run.diag = FALSE, sep = ",", silent = FALSE, ... )bfw( job.title = NULL, job.group = NULL, jags.model, jags.seed = NULL, jags.method = NULL, jags.chains = NULL, custom.function = NULL, custom.model = NULL, params = NULL, saved.steps = 10000, thinned.steps = 1, adapt.steps = NULL, burnin.steps = NULL, initial.list = list(), custom.name = NULL, project.name = "Project", project.dir = "Results/", project.data = NULL, time.stamp = TRUE, save.data = FALSE, data.set = "AllData", data.format = "csv", raw.data = FALSE, run.robust = FALSE, merge.MCMC = FALSE, run.diag = FALSE, sep = ",", silent = FALSE, ... )
job.title |
title of analysis, Default: NULL |
job.group |
for some hierarchical models with several layers of parameter names (e.g., latent and observed parameters), Default: NULL |
jags.model |
specify which module to use |
jags.seed |
specify seed to replicate a analysis, Default: NULL |
jags.method |
specify method for JAGS (e.g., parallel or simple), Default: NULL |
jags.chains |
specify specify number of chains for JAGS, Default: NULL |
custom.function |
custom function to use (e.g., defined function, external R file or string with function), Default: NULL |
custom.model |
define a custom model to use (e.g., string or text file (.txt), Default: NULL |
params |
define parameters to observe, Default: NULL |
saved.steps |
define the number of iterations/steps/chains in the MCMC simulations, Default: 10000 |
thinned.steps |
save every kth step of the original saved.steps, Default: 1 |
adapt.steps |
the number of adaptive iterations to use at the start of each simulation, Default: NULL |
burnin.steps |
the number of burnin iterations, NOT including the adaptive iterations to use for the simulation, Default: NULL |
initial.list |
initial values for analysis, Default: list() |
custom.name |
custom name of project, Default: NULL |
project.name |
name of project, Default: 'Project' |
project.dir |
define where to save data, Default: 'Results/' |
project.data |
define data to use for analysis (e.g., csv, rda, custom data.frame or matrix, or data included in package, Default: NULL |
time.stamp |
logical, indicating whether or not to append unix time stamp to file name, Default: TRUE |
save.data |
logical, indicating whether or not to save data, Default: FALSE |
data.set |
define subset of data, Default: 'AllData' |
data.format |
define what data format is being used, Default: 'csv' |
raw.data |
logical, indicating whether or not to use unprocessed data, Default: FALSE |
run.robust |
logical, indicating whether or not robust analysis, Default: FALSE |
merge.MCMC |
logical, indicating whether or not to merge MCMC chains, Default: FALSE |
run.diag |
logical, indicating whether or not to run diagnostics, Default: FALSE |
sep |
symbol to separate data (e.g., comma-delimited), Default: ',' |
silent |
logical, indicating whether or not to run analysis without output, Default: FALSE |
... |
further arguments passed to or from other methods |
Settings act like the main framework for bfw, connecting function, model and JAGS.
data from MCMC RunMCMC
head,modifyList,capture.output
capitalize the first letter in each words in a string
CapWords(s, strict = FALSE)CapWords(s, strict = FALSE)
s |
string |
strict |
logical, indicating whether or not string it set to title case , Default: FALSE |
returns capitalized string
CapWords("example eXAMPLE", FALSE) # [1] "Example EXAMPLE" CapWords("example eXAMPLE", TRUE) # [1] "Example Example"CapWords("example eXAMPLE", FALSE) # [1] "Example EXAMPLE" CapWords("example eXAMPLE", TRUE) # [1] "Example Example"
Shamelessly adapted from Field (2017).
CatsCats
A data frame with 2000 rows and 4 variables:
Rewardinteger Food or Affection
Danceinteger Yes or No
Alignmentinteger Good or Evil
Ratingsdouble Cats rate their owners (average of multiple seven-point Likert-type scale (1 = Hate ... 7 = Love)
Example data for BFW
Change names, colnames or rownames of single items or a list of items
ChangeNames( x, names, single.items = FALSE, row.names = FALSE, param = NULL, where = NULL, environment = NULL )ChangeNames( x, names, single.items = FALSE, row.names = FALSE, param = NULL, where = NULL, environment = NULL )
x |
list, vector, matrix, dataframe or a list of such items |
names |
names to insert |
single.items |
logical, indicating whether or not to use names rather than colnames or rownames, Default: FALSE |
row.names |
logical, indicating whether or not to use rownames rather than colnames, Default: FALSE |
param |
Variable name, Default: NULL |
where |
select parents, Default: NULL |
environment |
select reference environment, Default: NULL |
returns Named items # ABC <- c("1","2","3") # "1" "2" "3" # ChangeNames(ABC, names = c("A","B","C") , single.items = TRUE) # A B C # "1" "2" "3"
Compute highest density interval (HDI) from posterior output
ComputeHDI(data, credible.region)ComputeHDI(data, credible.region)
data |
data to compute HDI from |
credible.region |
summarize uncertainty by defining a region of most credible values (e.g., 95 percent of the distribution), Default: 0.95 |
values within the HDI have higher probability density than values outside the HDI, and the values inside the HDI have a total probability equal to the credible region (e.g., 95 percent).
Return HDI
set.seed(1) data <-rnorm(100,0,1) credible.region <- 0.95 ComputeHDI(data,credible.region) # HDIlo HDIhi # -1.99 1.60set.seed(1) data <-rnorm(100,0,1) credible.region <- 0.95 ComputeHDI(data,credible.region) # HDIlo HDIhi # -1.99 1.60
utilize the AddNames function to create contrast names
ContrastNames(items, job.names, col.names)ContrastNames(items, job.names, col.names)
items |
items to create names for |
job.names |
names of all parameters in analysis, Default: NULL |
col.names |
columns in MCMC to create names from |
MCMC convergence diagnostics
DiagMCMC( data.MCMC, par.name, job.names, job.group, credible.region = 0.95, monochrome = TRUE, plot.colors = c("#495054", "#e3e8ea") )DiagMCMC( data.MCMC, par.name, job.names, job.group, credible.region = 0.95, monochrome = TRUE, plot.colors = c("#495054", "#e3e8ea") )
data.MCMC |
MCMC chains to diagnose |
par.name |
parameter to analyze |
job.names |
names of all parameters in analysis, Default: NULL |
job.group |
for some hierarchical models with several layers of parameter names (e.g., latent and observed parameters), Default: NULL |
credible.region |
summarize uncertainty by defining a region of most credible values (e.g., 95 percent of the distribution), Default: 0.95 |
monochrome |
logical, indicating whether or not to use monochrome colors, else use DistinctColors, Default: TRUE |
plot.colors |
range of color to use, Default: c("#495054", "#e3e8ea") |
list of diagnostic plots
dev.new,colorRampPalette,recordPlot,graphics.off,dev.list,dev.off
par,layout,plot.new,matplot,abline,text,points,mtext
traceplot,gelman.plot,effectiveSize
sd,acf,density
create vector containing Hex color codes
DistinctColors(range, random = FALSE)DistinctColors(range, random = FALSE)
range |
number of colors as sequence |
random |
logical, indicating whether or not to provide random colors, Default: FALSE |
DistinctColors(1:3) # [1] "#FFFF00" "#1CE6FF" "#FF34FF" set.seed(1) DistinctColors(1:3, TRUE) # [1] "#575329" "#CB7E98" "#D86A78"DistinctColors(1:3) # [1] "#FFFF00" "#1CE6FF" "#FF34FF" set.seed(1) DistinctColors(1:3, TRUE) # [1] "#575329" "#CB7E98" "#D86A78"
Print estimated time for arrival (ETA)
ETA(start.time, i, total, results = NULL)ETA(start.time, i, total, results = NULL)
start.time |
start time (preset variable with Sys.time()) |
i |
incremental steps towards total |
total |
total number of steps |
results |
message to display, Default: NULL |
simple function to construct a file name for data
FileName( project = "Project", subset = NULL, type = NULL, name = NULL, unix = TRUE, ... )FileName( project = "Project", subset = NULL, type = NULL, name = NULL, unix = TRUE, ... )
project |
name of project, Default: 'Project' |
subset |
define subset of data, Default: NULL |
type |
type of data, Default: NULL |
name |
save name, Default: NULL |
unix |
logical, indicating whether or not to append unix timestamp, Default: TRUE |
... |
further arguments passed to or from other methods |
FileName() # [1] "Project-Name-1528834963" FileName(project = "Project" , subset = "subset" , type = "longitudinal" , name = "cheese", unix = FALSE) # [1] "Projectsubset-longitudinal-cheese"FileName() # [1] "Project-Name-1528834963" FileName(project = "Project" , subset = "subset" , type = "longitudinal" , name = "cheese", unix = FALSE) # [1] "Projectsubset-longitudinal-cheese"
Find the environment of a selected variable.
FindEnvironment(x, where = NULL)FindEnvironment(x, where = NULL)
x |
any type of named object |
where |
select reference environment, Default: NULL |
returns Found environment, Default: R_GlobalEnv.
flatten a nested list into a single list
FlattenList(li, rm.duplicated = TRUE, unname.li = TRUE, rm.empty = TRUE)FlattenList(li, rm.duplicated = TRUE, unname.li = TRUE, rm.empty = TRUE)
li |
list to flatten |
rm.duplicated |
logical, indicating whether or not to remove duplicated lists, Default: TRUE |
unname.li |
logical, indicating whether or not to unname lists, Default: TRUE |
rm.empty |
logical, indicating whether or not to remove empty lists, Default: TRUE |
li <- list(LETTERS[1:3], list(letters[1:3], list(LETTERS[4:6])), DEF = letters[4:6], LETTERS[1:3], list() # Emtpy list ) print(li) # [[1]] # [1] "A" "B" "C" # # [[2]] # [[2]][[1]] # [1] "a" "b" "c" # # [[2]][[2]] # [[2]][[2]][[1]] # [1] "D" "E" "F" # # # # $DEF # [1] "d" "e" "f" # # [[4]] # [1] "A" "B" "C" # # [[5]] # list() FlattenList(li) # [[1]] # [1] "A" "B" "C" # # [[2]] # [1] "a" "b" "c" # # [[3]] # [1] "D" "E" "F" # # [[4]] # [1] "d" "e" "f"li <- list(LETTERS[1:3], list(letters[1:3], list(LETTERS[4:6])), DEF = letters[4:6], LETTERS[1:3], list() # Emtpy list ) print(li) # [[1]] # [1] "A" "B" "C" # # [[2]] # [[2]][[1]] # [1] "a" "b" "c" # # [[2]][[2]] # [[2]][[2]][[1]] # [1] "D" "E" "F" # # # # $DEF # [1] "d" "e" "f" # # [[4]] # [1] "A" "B" "C" # # [[5]] # list() FlattenList(li) # [[1]] # [1] "A" "B" "C" # # [[2]] # [1] "a" "b" "c" # # [[3]] # [1] "D" "E" "F" # # [[4]] # [1] "d" "e" "f"
compute gamma distribution (shape and rate) from mode and standard deviation
GammaDist(mode, sd)GammaDist(mode, sd)
mode |
mode from data |
sd |
standard deviation from data |
GammaDist(1,0.5) # $shape # [1] 5.828427 # $rate # [1] 4.828427GammaDist(1,0.5) # $shape # [1] 5.828427 # $rate # [1] 4.828427
simple function to extract columns from data frame
GetRange(var, range = 1:8, df)GetRange(var, range = 1:8, df)
var |
variable of interest (e.g., V) |
range |
range of variables with same stem name (e.g., V1, V2, ..., V8) , Default: 1:8 |
df |
data to extract from |
data <- as.data.frame(matrix(1:80,ncol=8)) GetRange("V", c(1,4), data) # V1 V4 # 1 1 31 # 2 2 32 # 3 3 33 # 4 4 34 # 5 5 35 # 6 6 36 # 7 7 37 # 8 8 38 # 9 9 39 # 10 10 40data <- as.data.frame(matrix(1:80,ncol=8)) GetRange("V", c(1,4), data) # V1 V4 # 1 1 31 # 2 2 32 # 3 3 33 # 4 4 34 # 5 5 35 # 6 6 36 # 7 7 37 # 8 8 38 # 9 9 39 # 10 10 40
mix vectors by alternating between them
Interleave(a, b)Interleave(a, b)
a |
first vector |
b |
second vector |
mixed vector
a <- 1:3 b <- LETTERS[1:3] Interleave(a,b) # [1] "1" "A" "2" "B" "3" "C"a <- 1:3 b <- LETTERS[1:3] Interleave(a,b) # [1] "1" "A" "2" "B" "3" "C"
Compute inverse cumulative density function of the distribution
InverseHDI( beta, shape1, shape2, credible.region = 0.95, tolerance = 0.00000001 )InverseHDI( beta, shape1, shape2, credible.region = 0.95, tolerance = 0.00000001 )
beta |
density, distribution function, quantile function and random generation for the Beta distribution with parameters shape1 and shape2 |
shape1 |
non-negative parameter of the Beta distribution. |
shape2 |
non-negative parameter of the Beta distribution. |
credible.region |
summarize uncertainty by defining a region of most credible values (e.g., 95 percent of the distribution), Default: 0.95 |
tolerance |
the desired accuracy, Default: 1e-8 |
values within the HDI have higher probability density than values outside the HDI, and the values inside the HDI have a total probability equal to the credible region (e.g., 95 percent).
Return HDI
InverseHDI( qbeta , 554 , 149 ) # HDIlo HDIhi # 0.758 0.818InverseHDI( qbeta , 554 , 149 ) # HDIlo HDIhi # 0.758 0.818
collection of layout sizes
Layout(x = "a4", layout.inverse = FALSE)Layout(x = "a4", layout.inverse = FALSE)
x |
type of layout, Default: 'a4' |
layout.inverse |
logical, indicating whether or not to inverse layout (e.g., landscape) , Default: FALSE |
width and height of select medium
Layout() # [1] 8.3 11.7Layout() # [1] 8.3 11.7
Create matrices from combinations of columns
MatrixCombn( matrix, first.stem, last.stem = NULL, q.levels, rm.last = TRUE, row.means = TRUE )MatrixCombn( matrix, first.stem, last.stem = NULL, q.levels, rm.last = TRUE, row.means = TRUE )
matrix |
matrix to combine |
first.stem |
first name of columns to use (e.g., "m" for mean) |
last.stem |
optional last name of columns to use (e.g., "p" for proportions) , Default: NONE |
q.levels |
number of levels per column |
rm.last |
logical, indicating whether or not to remove last combination (i.e., m1m2m3m4) , Default: TRUE |
row.means |
logical, indicating whether or not to compute row means from combined columns, else use row sums, Default: TRUE |
Merge two or more MCMC simulations
MergeMCMC(pat, project.dir = "Results/", data.sets)MergeMCMC(pat, project.dir = "Results/", data.sets)
pat |
pattern to select MCMC chain from |
project.dir |
define where to save data, Default: 'Results/' |
data.sets |
data sets to combine |
Merged MCMC chains
Use multiple patterns from vector to find element in another vector, with option to remove certain patterns
MultiGrep(find, from, remove = NULL, value = TRUE)MultiGrep(find, from, remove = NULL, value = TRUE)
find |
vector to find |
from |
vector to find from |
remove |
variables to remove, Default: NULL |
value |
logical, if TRUE returns value, Default: TRUE |
simple function to normalize data
Normalize(x)Normalize(x)
x |
numeric vector to normalize |
Normalize(1:10) # [1] 0.0182 0.0364 0.0545 0.0727 0.0909 # 0.1091 0.1273 0.1455 0.1636 0.1818Normalize(1:10) # [1] 0.0182 0.0364 0.0545 0.0727 0.0909 # 0.1091 0.1273 0.1455 0.1636 0.1818
Pad a numeric vector according to the highest value
PadVector(v)PadVector(v)
v |
numeric vector to pad |
PadVector(1:10) # [1] "01" "02" "03" "04" "05" "06" "07" "08" "09" "10"PadVector(1:10) # [1] "01" "02" "03" "04" "05" "06" "07" "08" "09" "10"
simple function to extract numbers from string/vector
ParseNumber(x, digits = FALSE)ParseNumber(x, digits = FALSE)
x |
string or vector |
digits |
logical, indicating whether or not to extract decimals, Default: FALSE |
ParseNumber("String1WithNumbers2") # [1] 1 2ParseNumber("String1WithNumbers2") # [1] 1 2
Display and/or save plots
ParsePlot( plot.data, project.dir = "Results/", project.name = FileName(name = "Print"), graphic.type = "pdf", plot.size = "15,10", scaling = 100, plot.aspect = NULL, save.data = FALSE, vector.graphic = FALSE, point.size = 12, font.type = "serif", one.file = TRUE, ppi = 300, units = "in", layout = "a4", layout.inverse = FALSE, return.files = FALSE, ... )ParsePlot( plot.data, project.dir = "Results/", project.name = FileName(name = "Print"), graphic.type = "pdf", plot.size = "15,10", scaling = 100, plot.aspect = NULL, save.data = FALSE, vector.graphic = FALSE, point.size = 12, font.type = "serif", one.file = TRUE, ppi = 300, units = "in", layout = "a4", layout.inverse = FALSE, return.files = FALSE, ... )
plot.data |
a list of plots |
project.dir |
define where to save data, Default: 'Results/' |
project.name |
define name of project, Default: 'FileName(name="Print")' |
graphic.type |
type of graphics to use (e.g., pdf, png, ps), Default: 'pdf' |
plot.size |
size of plot, Default: '15,10' |
scaling |
scale size of plot, Default: 100 |
plot.aspect |
aspect of plot, Default: NULL |
save.data |
logical, indicating whether or not to save data, Default: FALSE |
vector.graphic |
logical, indicating whether or not visualizations should be vector or raster graphics, Default: FALSE |
point.size |
point size used for visualizations, Default: 12 |
font.type |
font type used for visualizations, Default: 'serif' |
one.file |
logical, indicating whether or not visualizations should be placed in one or several files, Default: TRUE |
ppi |
define pixel per inch used for visualizations, Default: 300 |
units |
define unit of length used for visualizations, Default: 'in' |
layout |
define a layout size for visualizations, Default: 'a4' |
layout.inverse |
logical, indicating whether or not to inverse layout (e.g., landscape) , Default: FALSE |
return.files |
logical, indicating whether or not to return saved file names |
... |
further arguments passed to or from other methods |
dev,
png,
ps.options,
recordPlot
head
readPNG
par,
plot,
rasterImage
read_pptx,
add_slide,
ph_with
dml
# Create three plots plot.data <- lapply(1:3, function (i) { # Open new device grDevices::dev.new() # Print plot plot(1:i) # Record plot p <- grDevices::recordPlot() # Turn off graphics device drive grDevices::dev.off() return (p) } ) # Print plots ParsePlot(plot.data)# Create three plots plot.data <- lapply(1:3, function (i) { # Open new device grDevices::dev.new() # Print plot plot(1:i) # Record plot p <- grDevices::recordPlot() # Turn off graphics device drive grDevices::dev.off() return (p) } ) # Print plots ParsePlot(plot.data)
Create a circlize plot
PlotCirclize( data, category.spacing = 1.2, category.inset = c(-0.4, 0), monochrome = TRUE, plot.colors = c("#CCCCCC", "#DEDEDE"), font.type = "serif" )PlotCirclize( data, category.spacing = 1.2, category.inset = c(-0.4, 0), monochrome = TRUE, plot.colors = c("#CCCCCC", "#DEDEDE"), font.type = "serif" )
data |
data for circlize plot |
category.spacing |
spacing between category items , Default: 1.25 |
category.inset |
inset of category items form plot , Default: c(-0.5, 0) |
monochrome |
logical, indicating whether or not to use monochrome colors, else use DistinctColors, Default: TRUE |
plot.colors |
range of color to use, Default: c("#CCCCCC", "#DEDEDE") |
font.type |
font type used for visualizations, Default: 'serif' |
dev,
recordPlot
legend
circos.par,
chordDiagram,
circos.trackPlotRegion,
circos.clear
Plot data as violin plot visualizing density, box plots to display HDI, whiskers to display standard deviation
PlotData(data, data.type = "Mean", ...)PlotData(data, data.type = "Mean", ...)
data |
data to plot data from |
data.type |
define what kind of data is being used, Default: 'Mean' |
... |
further arguments passed to or from other methods |
Create a (repeated) mean plot
PlotMean( data, monochrome = TRUE, plot.colors = c("#495054", "#e3e8ea"), font.type = "serif", run.repeated = FALSE, run.split = FALSE, y.split = FALSE, ribbon.plot = TRUE, y.text = "Score", x.text = NULL, remove.x = FALSE )PlotMean( data, monochrome = TRUE, plot.colors = c("#495054", "#e3e8ea"), font.type = "serif", run.repeated = FALSE, run.split = FALSE, y.split = FALSE, ribbon.plot = TRUE, y.text = "Score", x.text = NULL, remove.x = FALSE )
data |
MCMC data to plot |
monochrome |
logical, indicating whether or not to use monochrome colors, else use DistinctColors, Default: TRUE |
plot.colors |
range of color to use, Default: c("#495054", "#e3e8ea") |
font.type |
font type used for visualizations, Default: 'serif' |
run.repeated |
logical, indicating whether or not to use repeated measures plot, Default: FALSE |
run.split |
logical, indicating whether or not to use split violin plot and compare distribution between groups, Default: FALSE |
y.split |
logical, indicating whether or not to split within (TRUE) or between groups, Default: FALSE |
ribbon.plot |
logical, indicating whether or not to use ribbon plot for HDI, Default: TRUE |
y.text |
label on y axis, Default: 'Score' |
x.text |
label on x axis, Default: NULL |
remove.x |
logical, indicating whether or not to show x.axis information, Default: FALSE |
ggproto,
ggplot2-ggproto,
aes,
margin,
geom_boxplot,
geom_crossbar,
geom_path,
geom_ribbon,
geom_violin,
ggplot,
scale_manual,
scale_x_discrete,
theme,
layer,
labs
arrange,
rbind.fill
zero_range
grid.grob,
grobName,
unit
approxfun
colorRamp
Create a nominal plot
PlotNominal( data, monochrome = TRUE, plot.colors = c("#CCCCCC", "#DEDEDE"), font.type = "serif", bar.dodge = 0.6, bar.alpha = 0.7, bar.width = 0.4, bar.extras.dodge = 0, bar.border = "black", bar.label = FALSE, bar.error = TRUE, use.cutoff = FALSE, diff.cutoff = 1, q.items = NULL )PlotNominal( data, monochrome = TRUE, plot.colors = c("#CCCCCC", "#DEDEDE"), font.type = "serif", bar.dodge = 0.6, bar.alpha = 0.7, bar.width = 0.4, bar.extras.dodge = 0, bar.border = "black", bar.label = FALSE, bar.error = TRUE, use.cutoff = FALSE, diff.cutoff = 1, q.items = NULL )
data |
MCMC data to plot |
monochrome |
logical, indicating whether or not to use monochrome colors, else use DistinctColors, Default: TRUE |
plot.colors |
range of color to use, Default: c("#CCCCCC", "#DEDEDE") |
font.type |
font type used for visualizations, Default: 'serif' |
bar.dodge |
distance between within bar plots, Default: 0.6 |
bar.alpha |
transparency for bar plot, Default: 0.7 |
bar.width |
width of bar plot, Default: 0.4 |
bar.extras.dodge |
dodge of error bar and label, Default: 0 |
bar.border |
color of the bar border, Default: 'black' |
bar.label |
logical, indicating whether or not to show bar labels, Default: TRUE |
bar.error |
logical, indicating whether or not to show error bars, Default: TRUE |
use.cutoff |
logical, indicating whether or not to use a cutoff for keeping plots, Default: FALSE |
diff.cutoff |
if using a cutoff, determine the percentage that expected and observed values should differ, Default: 1 |
q.items |
which variables should be used in the plot. Defaults to all , Default: NULL |
aes,margin,geom_crossbar,ggplot,scale_manual,theme
Create a density plot with parameter values
PlotParam( data, param, ROPE = FALSE, monochrome = TRUE, plot.colors = c("#495054", "#e3e8ea"), font.type = "serif", font.size = 4.5, rope.line = -0.2, rope.tick = -0.1, rope.label = -0.35, line.size = 0.5, dens.zero.col = "black", dens.mean.col = "white", dens.median.col = "white", dens.mode.col = "black", dens.rope.col = "black", scale = FALSE, y.limits = NULL, y.breaks = NULL, x.limits = NULL, x.breaks = NULL, plot.title = NULL )PlotParam( data, param, ROPE = FALSE, monochrome = TRUE, plot.colors = c("#495054", "#e3e8ea"), font.type = "serif", font.size = 4.5, rope.line = -0.2, rope.tick = -0.1, rope.label = -0.35, line.size = 0.5, dens.zero.col = "black", dens.mean.col = "white", dens.median.col = "white", dens.mode.col = "black", dens.rope.col = "black", scale = FALSE, y.limits = NULL, y.breaks = NULL, x.limits = NULL, x.breaks = NULL, plot.title = NULL )
data |
MCMC data to plot |
param |
parameter of interest |
ROPE |
plot ROPE values, Default: FALSE |
monochrome |
logical, indicating whether or not to use monochrome colors, else use DistinctColors, Default: TRUE |
plot.colors |
range of color to use, Default: c("#495054", "#e3e8ea") |
font.type |
font type used for visualizations, Default: 'serif' |
font.size |
font size, Default: 4.5 |
rope.line |
size of ROPE lien, Default: -0.2 |
rope.tick |
distance to ROPE tick, Default: -0.1 |
rope.label |
distance to ROPE label, Default: -0.35 |
line.size |
overall line size, Default: 0.5 |
dens.zero.col |
colour of line indicating zero, Default: 'black' |
dens.mean.col |
colour of line indicating mean value, Default: 'white' |
dens.median.col |
colour of line indicating median value, Default: 'white' |
dens.mode.col |
colour of line indicating mode value, Default: 'black' |
dens.rope.col |
colour of line indicating ROPE value, Default: 'black' |
scale |
scale x and y axis, Default: FALSE |
y.limits |
vector of y limits, Default: NULL |
y.breaks |
vector of y breaks, Default: NULL |
x.limits |
= vector of x limits, Default: NULL |
x.breaks |
= vector of x breaks, Default: NULL |
plot.title |
= title of plot, Default: NULL |
Density plot of parameter values
mutate,group_by,join,select,slice,filter
approxfun
aes,margin,geom_density,geom_polygon,geom_segment,geom_label,ggplot,ggplot_build,scale_continuous,theme,labs
opens connection to a file
ReadFile( file = NULL, path = "models/", package = "bfw", type = "string", sep = ",", data.format = "txt", custom = FALSE )ReadFile( file = NULL, path = "models/", package = "bfw", type = "string", sep = ",", data.format = "txt", custom = FALSE )
file |
name of file, Default: NULL |
path |
path to file, Default: 'models/' |
package |
choose package to open from, Default: 'bfw' |
type |
Type of file (i.e., text or data), Default: 'string' |
sep |
symbol to separate data (e.g., comma-delimited), Default: ',' |
data.format |
define what data format is being used, Default: 'csv' |
custom |
logical, indicating whether or not to use custom file, , Default: FALSE |
# Print JAGS model for bernoulli trials cat(ReadFile("stats_bernoulli")) # model { # for (i in 1:n){ # x[i] ~ dbern(theta) # } # theta ~ dunif(0,1) # }# Print JAGS model for bernoulli trials cat(ReadFile("stats_bernoulli")) # model { # for (i in 1:n){ # x[i] ~ dbern(theta) # } # theta ~ dunif(0,1) # }
Remove empty elements in vector
RemoveEmpty(x)RemoveEmpty(x)
x |
vector to eliminate NA and blanks |
RemoveEmpty( c("",NA,"","Remains") ) # [1] "Remains"RemoveEmpty( c("",NA,"","Remains") ) # [1] "Remains"
Remove variable(s) and remove garbage from memory
RemoveGarbage(v)RemoveGarbage(v)
v |
variables to remove |
simple function to remove whitespace
RemoveSpaces(x)RemoveSpaces(x)
x |
string |
RemoveSpaces(" No More S p a c e s") # [1] "NoMoreSpaces"RemoveSpaces(" No More S p a c e s") # [1] "NoMoreSpaces"
Compute contrasts from mean and standard deviation (Cohen's d) or frequencies (odds ratio)
RunContrasts(contrast.type, q.levels, use.contrast, contrasts, data, job.names)RunContrasts(contrast.type, q.levels, use.contrast, contrasts, data, job.names)
contrast.type |
type of contrast: "m" indicate means and standard deviations, "o" indicate frequency |
q.levels |
Number of levels of each variable/column |
use.contrast |
choose from "between", "within" and "mixed". Between compare groups at different conditions. Within compare a group at different conditions. Mixed compute all comparisons |
contrasts |
specified contrasts columns |
data |
data to compute contrasts from |
job.names |
names of all parameters in analysis, Default: NULL |
Conduct MCMC simulations using JAGS
RunMCMC( jags.model, params = NULL, name.list, data.list, initial.list = list(), run.contrasts = FALSE, use.contrast = "between", contrasts = NULL, custom.contrast = NULL, run.ppp = FALSE, k.ppp = 10, n.data, credible.region = 0.95, save.data = FALSE, ROPE = NULL, merge.MCMC = FALSE, run.diag = FALSE, param.diag = NULL, sep = ",", monochrome = TRUE, plot.colors = c("#495054", "#e3e8ea"), graphic.type = "pdf", plot.size = "15,10", scaling = 100, plot.aspect = NULL, vector.graphic = FALSE, point.size = 12, font.type = "serif", one.file = TRUE, ppi = 300, units = "in", layout = "a4", layout.inverse = FALSE, ... )RunMCMC( jags.model, params = NULL, name.list, data.list, initial.list = list(), run.contrasts = FALSE, use.contrast = "between", contrasts = NULL, custom.contrast = NULL, run.ppp = FALSE, k.ppp = 10, n.data, credible.region = 0.95, save.data = FALSE, ROPE = NULL, merge.MCMC = FALSE, run.diag = FALSE, param.diag = NULL, sep = ",", monochrome = TRUE, plot.colors = c("#495054", "#e3e8ea"), graphic.type = "pdf", plot.size = "15,10", scaling = 100, plot.aspect = NULL, vector.graphic = FALSE, point.size = 12, font.type = "serif", one.file = TRUE, ppi = 300, units = "in", layout = "a4", layout.inverse = FALSE, ... )
jags.model |
specify which module to use |
params |
define parameters to observe, Default: NULL |
name.list |
list of names |
data.list |
list of data |
initial.list |
initial values for analysis, Default: list() |
run.contrasts |
logical, indicating whether or not to run contrasts, Default: FALSE |
use.contrast |
choose from "between", "within" and "mixed". Between compare groups at different conditions. Within compare a group at different conditions. Mixed compute all comparisons, Default: "between", |
contrasts |
define contrasts to use for analysis (defaults to all) , Default: NULL |
custom.contrast |
define contrasts for custom models , Default: NULL |
run.ppp |
logical, indicating whether or not to conduct ppp analysis, Default: FALSE |
k.ppp |
run ppp for every kth length of MCMC chains, Default: 10 |
n.data |
sample size for each parameter |
credible.region |
summarize uncertainty by defining a region of most credible values (e.g., 95 percent of the distribution), Default: 0.95 |
save.data |
logical, indicating whether or not to save data, Default: FALSE |
ROPE |
define range for region of practical equivalence (e.g., c(-0.05 , 0.05), Default: NULL |
merge.MCMC |
logical, indicating whether or not to merge MCMC chains, Default: FALSE |
run.diag |
logical, indicating whether or not to run diagnostics, Default: FALSE |
param.diag |
define parameters to use for diagnostics, default equals all parameters, Default: NULL |
sep |
symbol to separate data (e.g., comma-delimited), Default: ',' |
monochrome |
logical, indicating whether or not to use monochrome colors, else use DistinctColors, Default: TRUE |
plot.colors |
range of color to use, Default: c("#495054", "#e3e8ea") |
graphic.type |
type of graphics to use (e.g., pdf, png, ps), Default: 'pdf' |
plot.size |
size of plot, Default: '15,10' |
scaling |
scale size of plot, Default: 100 |
plot.aspect |
aspect of plot, Default: NULL |
vector.graphic |
logical, indicating whether or not visualizations should be vector or raster graphics, Default: FALSE |
point.size |
point size used for visualizations, Default: 12 |
font.type |
font type used for visualizations, Default: 'serif' |
one.file |
logical, indicating whether or not visualizations should be placed in one or several files, Default: TRUE |
ppi |
define pixel per inch used for visualizations, Default: 300 |
units |
define unit of length used for visualizations, Default: 'in' |
layout |
define a layout size for visualizations, Default: 'a4' |
layout.inverse |
logical, indicating whether or not to inverse layout (e.g., landscape) , Default: FALSE |
... |
further arguments passed to or from other methods |
list containing MCMC chains , MCMC chains as matrix , summary of MCMC, list of name used, list of data, the jags model, running time of analysis and names of saved files
runjags.options,run.jags
detectCores
as.mcmc.list,varnames
rbind.fill
cor,cov,sd
mvrnorm
write.table
determine whether input is a single string
SingleString(x)SingleString(x)
x |
string |
true or false
A <- "This is a single string" SingleString(A) # [1] TRUE is.character(A) # [1] TRUE B <- c("This is a vector" , "containing two strings") SingleString(B) # [1] FALSE is.character(B) # [1] TRUEA <- "This is a single string" SingleString(A) # [1] TRUE is.character(A) # [1] TRUE B <- c("This is a vector" , "containing two strings") SingleString(B) # [1] FALSE is.character(B) # [1] TRUE
Conduct bernoulli trials
StatsBernoulli( x = NULL, x.names = NULL, DF, params = NULL, initial.list = list(), ... )StatsBernoulli( x = NULL, x.names = NULL, DF, params = NULL, initial.list = list(), ... )
x |
predictor variable(s), Default: NULL |
x.names |
optional names for predictor variable(s), Default: NULL |
DF |
data for analysis |
params |
define parameters to observe, Default: NULL |
initial.list |
initial values for analysis, Default: list() |
... |
further arguments passed to or from other methods |
## Create coin toss data: heads = 50 and tails = 50 #fair.coin<- as.matrix(c(rep("Heads",50),rep("Tails",50))) #colnames(fair.coin) <- "X" #fair.coin <- bfw(project.data = fair.coin, # x = "X", # saved.steps = 50000, # jags.model = "bernoulli", # jags.seed = 100, # ROPE = c(0.4,0.6), # silent = TRUE) #fair.coin.freq <- binom.test( 50000 * 0.5, 50000) ## Create coin toss data: heads = 20 and tails = 80 #biased.coin <- as.matrix(c(rep("Heads",20),rep("Tails",80))) #colnames(biased.coin) <- "X" #biased.coin <- bfw(project.data = biased.coin, # x = "X", # saved.steps = 50000, # jags.model = "bernoulli", # jags.seed = 101, # initial.list = list(theta = 0.7), # ROPE = c(0.4,0.6), # silent = TRUE) #biased.coin.freq <- binom.test( 50000 * 0.8, 50000) ## Print Bayesian and frequentist results of fair coin #fair.coin$summary.MCMC[,c(3:6,9:12)] ## Mode ESS HDIlo HDIhi ROPElo ROPEhi ROPEin n ## 0.505 50480.000 0.405 0.597 2.070 2.044 95.886 100.00 #sprintf("Frequentist: %.3f [%.3f , %.3f], p = %.3f" , # fair.coin.freq$estimate , # fair.coin.freq$conf.int[1] , # fair.coin.freq$conf.int[2] , # fair.coin.freq$p.value) ## [1] "Frequentist: 0.500 [0.496 , 0.504], p = 1.000" ## Print Bayesian and frequentist results of biased coin #biased.coin$summary.MCMC[,c(3:6,9:12)] ## Mode ESS HDIlo HDIhi ROPElo ROPEhi ROPEin n ## 0.803 50000.000 0.715 0.870 0.000 99.996 0.004 100.000 #sprintf("Frequentist: %.3f [%.3f , %.3f], p = %.3f" , # biased.coin.freq$estimate , # biased.coin.freq$conf.int[1] , # biased.coin.freq$conf.int[2] , # biased.coin.freq$p.value) ## [1] "Frequentist: 0.800 [0.796 , 0.803], p = 0.000"## Create coin toss data: heads = 50 and tails = 50 #fair.coin<- as.matrix(c(rep("Heads",50),rep("Tails",50))) #colnames(fair.coin) <- "X" #fair.coin <- bfw(project.data = fair.coin, # x = "X", # saved.steps = 50000, # jags.model = "bernoulli", # jags.seed = 100, # ROPE = c(0.4,0.6), # silent = TRUE) #fair.coin.freq <- binom.test( 50000 * 0.5, 50000) ## Create coin toss data: heads = 20 and tails = 80 #biased.coin <- as.matrix(c(rep("Heads",20),rep("Tails",80))) #colnames(biased.coin) <- "X" #biased.coin <- bfw(project.data = biased.coin, # x = "X", # saved.steps = 50000, # jags.model = "bernoulli", # jags.seed = 101, # initial.list = list(theta = 0.7), # ROPE = c(0.4,0.6), # silent = TRUE) #biased.coin.freq <- binom.test( 50000 * 0.8, 50000) ## Print Bayesian and frequentist results of fair coin #fair.coin$summary.MCMC[,c(3:6,9:12)] ## Mode ESS HDIlo HDIhi ROPElo ROPEhi ROPEin n ## 0.505 50480.000 0.405 0.597 2.070 2.044 95.886 100.00 #sprintf("Frequentist: %.3f [%.3f , %.3f], p = %.3f" , # fair.coin.freq$estimate , # fair.coin.freq$conf.int[1] , # fair.coin.freq$conf.int[2] , # fair.coin.freq$p.value) ## [1] "Frequentist: 0.500 [0.496 , 0.504], p = 1.000" ## Print Bayesian and frequentist results of biased coin #biased.coin$summary.MCMC[,c(3:6,9:12)] ## Mode ESS HDIlo HDIhi ROPElo ROPEhi ROPEin n ## 0.803 50000.000 0.715 0.870 0.000 99.996 0.004 100.000 #sprintf("Frequentist: %.3f [%.3f , %.3f], p = %.3f" , # biased.coin.freq$estimate , # biased.coin.freq$conf.int[1] , # biased.coin.freq$conf.int[2] , # biased.coin.freq$p.value) ## [1] "Frequentist: 0.800 [0.796 , 0.803], p = 0.000"
Covariate estimations (including correlation and Cronbach's alpha)
StatsCovariate( y = NULL, y.names = NULL, x = NULL, x.names = NULL, DF, params = NULL, job.group = NULL, initial.list = list(), jags.model, ... )StatsCovariate( y = NULL, y.names = NULL, x = NULL, x.names = NULL, DF, params = NULL, job.group = NULL, initial.list = list(), jags.model, ... )
y |
criterion variable(s), Default: NULL |
y.names |
optional names for criterion variable(s), Default: NULL |
x |
predictor variable(s), Default: NULL |
x.names |
optional names for predictor variable(s), Default: NULL |
DF |
data to analyze |
params |
define parameters to observe, Default: NULL |
job.group |
for some hierarchical models with several layers of parameter names (e.g., latent and observed parameters), Default: NULL |
initial.list |
initial values for analysis, Default: list() |
jags.model |
specify which module to use |
... |
further arguments passed to or from other methods |
covariate, correlation and (optional) Cronbach's alpha
## Create normal distributed data with mean = 0 and standard deviation = 1 ### r = 0.5 #data <- MASS::mvrnorm(n=100, # mu=c(0, 0), # Sigma=matrix(c(1, 0.5, 0.5, 1), 2), # empirical=TRUE) ## Add names #colnames(data) <- c("X","Y") ## Create noise with mean = 10 / -10 and sd = 1 ### r = -1.0 #noise <- MASS::mvrnorm(n=2, # mu=c(10, -10), # Sigma=matrix(c(1, -1, -1, 1), 2), # empirical=TRUE) ## Combine noise and data #biased.data <- rbind(data,noise) # # ## Run analysis on normal distributed data #mcmc <- bfw(project.data = data, # y = "X,Y", # saved.steps = 50000, # jags.model = "covariate", # jags.seed = 100, # silent = TRUE) ## Run robust analysis on normal distributed data #mcmc.robust <- bfw(project.data = data, # y = "X,Y", # saved.steps = 50000, # jags.model = "covariate", # run.robust = TRUE, # jags.seed = 101, # silent = TRUE) ## Run analysis on data with outliers #biased.mcmc <- bfw(project.data = biased.data, # y = "X,Y", # saved.steps = 50000, # jags.model = "covariate", # jags.seed = 102, # silent = TRUE) ## Run robust analysis on data with outliers #biased.mcmc.robust <- bfw(project.data = biased.data, # y = "X,Y", # saved.steps = 50000, # jags.model = "covariate", # run.robust = TRUE, # jags.seed = 103, # silent = TRUE) ## Print frequentist results #stats::cor(data)[2] ## [1] 0.5 #stats::cor(noise)[2] ## [1] -1 #stats::cor(biased.data)[2] ## [1] -0.498 ## Print Bayesian results #mcmc$summary.MCMC ## Mean Median Mode ESS HDIlo HDIhi n ## cor[1,1]: X vs. X 1.000 1.000 0.999 0 1.000 1.000 100 ## cor[2,1]: Y vs. X 0.488 0.491 0.496 19411 0.337 0.633 100 ## cor[1,2]: X vs. Y 0.488 0.491 0.496 19411 0.337 0.633 100 ## cor[2,2]: Y vs. Y 1.000 1.000 0.999 0 1.000 1.000 100 #mcmc.robust$summary.MCMC ## Mean Median Mode ESS HDIlo HDIhi n ## cor[1,1]: X vs. X 1.00 1.000 0.999 0 1.000 1.000 100 ## cor[2,1]: Y vs. X 0.47 0.474 0.491 18626 0.311 0.626 100 ## cor[1,2]: X vs. Y 0.47 0.474 0.491 18626 0.311 0.626 100 ## cor[2,2]: Y vs. Y 1.00 1.000 0.999 0 1.000 1.000 100 #biased.mcmc$summary.MCMC ## Mean Median Mode ESS HDIlo HDIhi n ## cor[1,1]: X vs. X 1.000 1.000 0.999 0 1.000 1.000 102 ## cor[2,1]: Y vs. X -0.486 -0.489 -0.505 19340 -0.627 -0.335 102 ## cor[1,2]: X vs. Y -0.486 -0.489 -0.505 19340 -0.627 -0.335 102 ## cor[2,2]: Y vs. Y 1.000 1.000 0.999 0 1.000 1.000 102 #biased.mcmc.robust$summary.MCMC ## Mean Median Mode ESS HDIlo HDIhi n ## cor[1,1]: X vs. X 1.000 1.000 0.999 0 1.000 1.000 102 ## cor[2,1]: Y vs. X 0.338 0.343 0.356 23450 0.125 0.538 102 ## cor[1,2]: X vs. Y 0.338 0.343 0.356 23450 0.125 0.538 102## Create normal distributed data with mean = 0 and standard deviation = 1 ### r = 0.5 #data <- MASS::mvrnorm(n=100, # mu=c(0, 0), # Sigma=matrix(c(1, 0.5, 0.5, 1), 2), # empirical=TRUE) ## Add names #colnames(data) <- c("X","Y") ## Create noise with mean = 10 / -10 and sd = 1 ### r = -1.0 #noise <- MASS::mvrnorm(n=2, # mu=c(10, -10), # Sigma=matrix(c(1, -1, -1, 1), 2), # empirical=TRUE) ## Combine noise and data #biased.data <- rbind(data,noise) # # ## Run analysis on normal distributed data #mcmc <- bfw(project.data = data, # y = "X,Y", # saved.steps = 50000, # jags.model = "covariate", # jags.seed = 100, # silent = TRUE) ## Run robust analysis on normal distributed data #mcmc.robust <- bfw(project.data = data, # y = "X,Y", # saved.steps = 50000, # jags.model = "covariate", # run.robust = TRUE, # jags.seed = 101, # silent = TRUE) ## Run analysis on data with outliers #biased.mcmc <- bfw(project.data = biased.data, # y = "X,Y", # saved.steps = 50000, # jags.model = "covariate", # jags.seed = 102, # silent = TRUE) ## Run robust analysis on data with outliers #biased.mcmc.robust <- bfw(project.data = biased.data, # y = "X,Y", # saved.steps = 50000, # jags.model = "covariate", # run.robust = TRUE, # jags.seed = 103, # silent = TRUE) ## Print frequentist results #stats::cor(data)[2] ## [1] 0.5 #stats::cor(noise)[2] ## [1] -1 #stats::cor(biased.data)[2] ## [1] -0.498 ## Print Bayesian results #mcmc$summary.MCMC ## Mean Median Mode ESS HDIlo HDIhi n ## cor[1,1]: X vs. X 1.000 1.000 0.999 0 1.000 1.000 100 ## cor[2,1]: Y vs. X 0.488 0.491 0.496 19411 0.337 0.633 100 ## cor[1,2]: X vs. Y 0.488 0.491 0.496 19411 0.337 0.633 100 ## cor[2,2]: Y vs. Y 1.000 1.000 0.999 0 1.000 1.000 100 #mcmc.robust$summary.MCMC ## Mean Median Mode ESS HDIlo HDIhi n ## cor[1,1]: X vs. X 1.00 1.000 0.999 0 1.000 1.000 100 ## cor[2,1]: Y vs. X 0.47 0.474 0.491 18626 0.311 0.626 100 ## cor[1,2]: X vs. Y 0.47 0.474 0.491 18626 0.311 0.626 100 ## cor[2,2]: Y vs. Y 1.00 1.000 0.999 0 1.000 1.000 100 #biased.mcmc$summary.MCMC ## Mean Median Mode ESS HDIlo HDIhi n ## cor[1,1]: X vs. X 1.000 1.000 0.999 0 1.000 1.000 102 ## cor[2,1]: Y vs. X -0.486 -0.489 -0.505 19340 -0.627 -0.335 102 ## cor[1,2]: X vs. Y -0.486 -0.489 -0.505 19340 -0.627 -0.335 102 ## cor[2,2]: Y vs. Y 1.000 1.000 0.999 0 1.000 1.000 102 #biased.mcmc.robust$summary.MCMC ## Mean Median Mode ESS HDIlo HDIhi n ## cor[1,1]: X vs. X 1.000 1.000 0.999 0 1.000 1.000 102 ## cor[2,1]: Y vs. X 0.338 0.343 0.356 23450 0.125 0.538 102 ## cor[1,2]: X vs. Y 0.338 0.343 0.356 23450 0.125 0.538 102
Apply latent or observed models to fit data (e.g., SEM, CFA, mediation)
StatsFit( latent = NULL, latent.names = NULL, observed = NULL, observed.names = NULL, additional = NULL, additional.names = NULL, DF, params = NULL, job.group = NULL, initial.list = list(), model.name, jags.model, custom.model = NULL, run.ppp = FALSE, run.robust = FALSE, ... )StatsFit( latent = NULL, latent.names = NULL, observed = NULL, observed.names = NULL, additional = NULL, additional.names = NULL, DF, params = NULL, job.group = NULL, initial.list = list(), model.name, jags.model, custom.model = NULL, run.ppp = FALSE, run.robust = FALSE, ... )
latent |
latenr variables, Default: NULL |
latent.names |
optional names for for latent variables, Default: NULL |
observed |
observed variable(s), Default: NULL |
observed.names |
optional names for for observed variable(s), Default: NULL |
additional |
supplemental parameters for fitted data (e.g., indirect pathways and total effect), Default: NULL |
additional.names |
optional names for supplemental parameters, Default: NULL |
DF |
data to analyze |
params |
define parameters to observe, Default: NULL |
job.group |
for some hierarchical models with several layers of parameter names (e.g., latent and observed parameters), Default: NULL |
initial.list |
initial values for analysis, Default: list() |
model.name |
name of model used |
jags.model |
specify which module to use |
custom.model |
define a custom model to use (e.g., string or text file (.txt), Default: NULL |
run.ppp |
logical, indicating whether or not to conduct ppp analysis, Default: FALSE |
run.robust |
logical, indicating whether or not robust analysis, Default: FALSE |
... |
further arguments passed to or from other methods |
Bayesian alternative to Cohen's kappa
StatsKappa( x = NULL, x.names = NULL, DF, params = NULL, initial.list = list(), ... )StatsKappa( x = NULL, x.names = NULL, DF, params = NULL, initial.list = list(), ... )
x |
predictor variable(s), Default: NULL |
x.names |
optional names for predictor variable(s), Default: NULL |
DF |
data to analyze |
params |
define parameters to observe, Default: NULL |
initial.list |
initial values for analysis, Default: list() |
... |
further arguments passed to or from other methods |
## Simulate rater data #Rater1 <- c(rep(0,20),rep(1,80)) #set.seed(100) #Rater2 <- c(rbinom(20,1,0.1), rbinom(80,1,0.9)) #data <- data.frame(Rater1,Rater2) #mcmc <- bfw(project.data = data, # x = "Rater1,Rater2", # saved.steps = 50000, # jags.model = "kappa", # jags.seed = 100, # silent = TRUE) ## Print frequentist and Bayesian kappa #library(psych) #psych::cohen.kappa(data)$confid[1,] ## lower estimate upper ## 0.6137906 0.7593583 0.9049260 ##' mcmc$summary.MCMC ## Mean Median Mode ESS HDIlo HDIhi n ## Kappa[1]: 0.739176 0.7472905 0.7634503 50657 0.578132 0.886647 100## Simulate rater data #Rater1 <- c(rep(0,20),rep(1,80)) #set.seed(100) #Rater2 <- c(rbinom(20,1,0.1), rbinom(80,1,0.9)) #data <- data.frame(Rater1,Rater2) #mcmc <- bfw(project.data = data, # x = "Rater1,Rater2", # saved.steps = 50000, # jags.model = "kappa", # jags.seed = 100, # silent = TRUE) ## Print frequentist and Bayesian kappa #library(psych) #psych::cohen.kappa(data)$confid[1,] ## lower estimate upper ## 0.6137906 0.7593583 0.9049260 ##' mcmc$summary.MCMC ## Mean Median Mode ESS HDIlo HDIhi n ## Kappa[1]: 0.739176 0.7472905 0.7634503 50657 0.578132 0.886647 100
Compute means and standard deviations.
StatsMean( y = NULL, y.names = NULL, x = NULL, x.names = NULL, DF, params = NULL, initial.list = list(), ... )StatsMean( y = NULL, y.names = NULL, x = NULL, x.names = NULL, DF, params = NULL, initial.list = list(), ... )
y |
criterion variable(s), Default: NULL |
y.names |
optional names for criterion variable(s), Default: NULL |
x |
categorical variable(s), Default: NULL |
x.names |
optional names for categorical variable(s), Default: NULL |
DF |
User defined data frame, Default: NULL |
params |
define parameters to observe, Default: NULL |
initial.list |
Initial values for simulations, Default: list() |
... |
further arguments passed to or from other methods |
mean and standard deviation
Bayesian alternative to ANOVA
StatsMetric( y = NULL, y.names = NULL, x = NULL, x.names = NULL, DF, params = NULL, job.group = NULL, initial.list = list(), model.name, jags.model, custom.model = NULL, run.robust = FALSE, ... )StatsMetric( y = NULL, y.names = NULL, x = NULL, x.names = NULL, DF, params = NULL, job.group = NULL, initial.list = list(), model.name, jags.model, custom.model = NULL, run.robust = FALSE, ... )
y |
criterion variable(s), Default: NULL |
y.names |
optional names for criterion variable(s), Default: NULL |
x |
categorical variable(s), Default: NULL |
x.names |
optional names for categorical variable(s), Default: NULL |
DF |
data to analyze |
params |
define parameters to observe, Default: NULL |
job.group |
for some hierarchical models with several layers of parameter names (e.g., latent and observed parameters), Default: NULL |
initial.list |
initial values for analysis, Default: list() |
model.name |
name of model used |
jags.model |
specify which module to use |
custom.model |
define a custom model to use (e.g., string or text file (.txt), Default: NULL |
run.robust |
logical, indicating whether or not robust analysis, Default: FALSE |
... |
further arguments passed to or from other methods |
complete.cases,
sd,
aggregate,
median
head
Bayesian alternative to chi-square test
StatsNominal( x = NULL, x.names = NULL, DF, params = NULL, job.group = NULL, initial.list = list(), model.name, jags.model, custom.model = NULL, ... )StatsNominal( x = NULL, x.names = NULL, DF, params = NULL, job.group = NULL, initial.list = list(), model.name, jags.model, custom.model = NULL, ... )
x |
categorical variable(s), Default: NULL |
x.names |
optional names for categorical variable(s), Default: NULL |
DF |
data to analyze |
params |
define parameters to observe, Default: NULL |
job.group |
for some hierarchical models with several layers of parameter names (e.g., latent and observed parameters), Default: NULL |
initial.list |
initial values for analysis, Default: list() |
model.name |
name of model used |
jags.model |
specify which module to use |
custom.model |
define a custom model to use (e.g., string or text file (.txt), Default: NULL |
... |
further arguments passed to or from other methods |
## Use cats data # mcmc <- bfw(project.data = bfw::Cats, # x = "Reward,Dance,Alignment", # saved.steps = 50000, # jags.model = "nominal", # run.contrasts = TRUE, # jags.seed = 100) ## Print only odds-ratio and effect sizes # mcmc$summary.MCMC[ grep("Odds ratio|Effect", # rownames(mcmc$summary.MCMC)) , c(3:7) ] ## Mode ESS HDIlo HDIhi n ## Effect size: Affection/Food vs. Evil/Good 0.12844 45222 0.00115 0.25510 2000 ## Effect size: Affection/Food vs. No/Yes 1.05346 44304 0.90825 1.18519 2000 ## Effect size: Affection/Food vs. No/Yes @ Evil 2.58578 30734 2.35471 2.85450 1299 ## Effect size: Affection/Food vs. No/Yes @ Good -0.51934 35316 -0.73443 -0.30726 701 ## Effect size: Food/Affection vs. Evil/Good -0.12844 45222 -0.25510 -0.00115 2000 ## Effect size: Food/Affection vs. No/Yes -1.05346 44304 -1.18519 -0.90825 2000 ## Effect size: Food/Affection vs. No/Yes @ Evil -2.58578 30734 -2.85450 -2.35471 1299 ## Effect size: Food/Affection vs. No/Yes @ Good 0.51934 35316 0.30726 0.73443 701 ## Effect size: No/Yes vs. Evil/Good 1.43361 43603 1.30715 1.55020 2000 ## Effect size: Yes/No vs. Evil/Good -1.43361 43603 -1.55020 -1.30715 2000 ## Odds ratio: Affection/Food vs. Evil/Good 1.25432 45225 0.99311 1.57765 2000 ## Odds ratio: Affection/Food vs. No/Yes 6.49442 44215 5.10392 8.46668 2000 ## Odds ratio: Affection/Food vs. No/Yes @ Evil 104.20109 30523 66.55346 169.12331 1299 ## Odds ratio: Affection/Food vs. No/Yes @ Good 0.36685 35417 0.25478 0.55982 701 ## Odds ratio: Food/Affection vs. Evil/Good 0.77604 45245 0.62328 0.98904 2000 ## Odds ratio: Food/Affection vs. No/Yes 0.14586 44452 0.11426 0.18982 2000 ## Odds ratio: Food/Affection vs. No/Yes @ Evil 0.00848 31117 0.00527 0.01336 1299 ## Odds ratio: Food/Affection vs. No/Yes @ Good 2.44193 35397 1.65204 3.63743 701 ## Odds ratio: No/Yes vs. Evil/Good 13.12995 43500 10.58859 16.49207 2000 ## Odds ratio: Yes/No vs. Evil/Good 0.07393 43739 0.05909 0.09221 2000 # ## The results indicate that evil cats are 13.13 times more likely than good cats to decline dancing ## Furthermore, when offered affection, evil cats are 104.20 times more likely to decline dancing, ## relative to evil cats that are offered food.## Use cats data # mcmc <- bfw(project.data = bfw::Cats, # x = "Reward,Dance,Alignment", # saved.steps = 50000, # jags.model = "nominal", # run.contrasts = TRUE, # jags.seed = 100) ## Print only odds-ratio and effect sizes # mcmc$summary.MCMC[ grep("Odds ratio|Effect", # rownames(mcmc$summary.MCMC)) , c(3:7) ] ## Mode ESS HDIlo HDIhi n ## Effect size: Affection/Food vs. Evil/Good 0.12844 45222 0.00115 0.25510 2000 ## Effect size: Affection/Food vs. No/Yes 1.05346 44304 0.90825 1.18519 2000 ## Effect size: Affection/Food vs. No/Yes @ Evil 2.58578 30734 2.35471 2.85450 1299 ## Effect size: Affection/Food vs. No/Yes @ Good -0.51934 35316 -0.73443 -0.30726 701 ## Effect size: Food/Affection vs. Evil/Good -0.12844 45222 -0.25510 -0.00115 2000 ## Effect size: Food/Affection vs. No/Yes -1.05346 44304 -1.18519 -0.90825 2000 ## Effect size: Food/Affection vs. No/Yes @ Evil -2.58578 30734 -2.85450 -2.35471 1299 ## Effect size: Food/Affection vs. No/Yes @ Good 0.51934 35316 0.30726 0.73443 701 ## Effect size: No/Yes vs. Evil/Good 1.43361 43603 1.30715 1.55020 2000 ## Effect size: Yes/No vs. Evil/Good -1.43361 43603 -1.55020 -1.30715 2000 ## Odds ratio: Affection/Food vs. Evil/Good 1.25432 45225 0.99311 1.57765 2000 ## Odds ratio: Affection/Food vs. No/Yes 6.49442 44215 5.10392 8.46668 2000 ## Odds ratio: Affection/Food vs. No/Yes @ Evil 104.20109 30523 66.55346 169.12331 1299 ## Odds ratio: Affection/Food vs. No/Yes @ Good 0.36685 35417 0.25478 0.55982 701 ## Odds ratio: Food/Affection vs. Evil/Good 0.77604 45245 0.62328 0.98904 2000 ## Odds ratio: Food/Affection vs. No/Yes 0.14586 44452 0.11426 0.18982 2000 ## Odds ratio: Food/Affection vs. No/Yes @ Evil 0.00848 31117 0.00527 0.01336 1299 ## Odds ratio: Food/Affection vs. No/Yes @ Good 2.44193 35397 1.65204 3.63743 701 ## Odds ratio: No/Yes vs. Evil/Good 13.12995 43500 10.58859 16.49207 2000 ## Odds ratio: Yes/No vs. Evil/Good 0.07393 43739 0.05909 0.09221 2000 # ## The results indicate that evil cats are 13.13 times more likely than good cats to decline dancing ## Furthermore, when offered affection, evil cats are 104.20 times more likely to decline dancing, ## relative to evil cats that are offered food.
Simple, multiple and hierarchical regression
StatsRegression( y = NULL, y.names = NULL, x = NULL, x.names = NULL, x.steps = NULL, x.blocks = NULL, DF, params = NULL, job.group = NULL, initial.list = list(), ... )StatsRegression( y = NULL, y.names = NULL, x = NULL, x.names = NULL, x.steps = NULL, x.blocks = NULL, DF, params = NULL, job.group = NULL, initial.list = list(), ... )
y |
criterion variable(s), Default: NULL |
y.names |
optional names for criterion variable(s), Default: NULL |
x |
predictor variable(s), Default: NULL |
x.names |
optional names for predictor variable(s), Default: NULL |
x.steps |
define number of steps in hierarchical regression , Default: NULL |
x.blocks |
define which predictors are included in each step (e.g., for three steps "1,2,3") , Default: NULL |
DF |
data to analyze |
params |
define parameters to observe, Default: NULL |
job.group |
for some hierarchical models with several layers of parameter names (e.g., latent and observed parameters), Default: NULL |
initial.list |
initial values for analysis, Default: list() |
... |
further arguments passed to or from other methods |
Perform softmax regression (i.e., multinomial logistic regression)
StatsSoftmax( y = NULL, y.names = NULL, x = NULL, x.names = NULL, DF, params = NULL, job.group = NULL, initial.list = NULL, run.robust = FALSE, ... )StatsSoftmax( y = NULL, y.names = NULL, x = NULL, x.names = NULL, DF, params = NULL, job.group = NULL, initial.list = NULL, run.robust = FALSE, ... )
y |
criterion variable(s), Default: NULL |
y.names |
optional names for criterion variable(s), Default: NULL |
x |
predictor variable(s), Default: NULL |
x.names |
optional names for predictor variable(s), Default: NULL |
DF |
data to analyze |
params |
define parameters to observe, Default: NULL |
job.group |
for some hierarchical models with several layers of parameter names (e.g., latent and observed parameters), Default: NULL |
initial.list |
initial values for analysis, Default: list() |
run.robust |
logical, indicating whether or not robust analysis, Default: FALSE |
... |
further arguments passed to or from other methods |
## Conduct softmax regression on Cats data ### Reward is 0 = Food and 1 = Dance ### Sample 100 datapoints from Cats data #mcmc <- bfw(project.data = bfw::Cats, # y = "Alignment", # x = "Ratings,Reward", # saved.steps = 50000, # jags.model = "softmax", # jags.seed = 100) ## Conduct binominal generalized linear model #model <- glm(Alignment ~ Ratings + Reward, data=bfw::Cats, family = binomial(link="logit")) ## Print output from softmax #mcmc$summary.MCMC # ## Mean Median Mode ESS HDIlo HDIhi n ##beta[1,1]: Evil vs. Ratings 0.000 0.00 -0.000607 0 0.000 0.000 2000 ##beta[1,2]: Evil vs. Reward 0.000 0.00 -0.000607 0 0.000 0.000 2000 ##beta[2,1]: Good vs. Ratings 1.289 1.29 1.283403 19614 1.187 1.387 2000 ##beta[2,2]: Good vs. Reward 1.276 1.27 1.279209 20807 0.961 1.597 2000 ##beta0[1]: Intercept: Evil 0.000 0.00 -0.000607 0 0.000 0.000 2000 ##beta0[2]: Intercept: Good -7.690 -7.68 -7.659198 17693 -8.472 -6.918 2000 ##zbeta[1,1]: Evil vs. Ratings 0.000 0.00 -0.000607 0 0.000 0.000 2000 ##zbeta[1,2]: Evil vs. Reward 0.000 0.00 -0.000607 0 0.000 0.000 2000 ##zbeta[2,1]: Good vs. Ratings 2.476 2.47 2.464586 19614 2.280 2.664 2000 ##zbeta[2,2]: Good vs. Reward 0.501 0.50 0.501960 20807 0.377 0.626 2000 ##zbeta0[1]: Intercept: Evil 0.000 0.00 -0.000607 0 0.000 0.000 2000 ##zbeta0[2]: Intercept: Good -1.031 -1.03 -1.024178 22812 -1.185 -0.870 2000 # ## Print (truncated) output from GML ## Estimate Std. Error z value Pr(>|z|) ##(Intercept) -6.39328 0.27255 -23.457 < 2e-16 *** ##Ratings 1.28480 0.05136 25.014 < 2e-16 *** ##RewardAffection 1.26975 0.16381 7.751 9.1e-15 ***## Conduct softmax regression on Cats data ### Reward is 0 = Food and 1 = Dance ### Sample 100 datapoints from Cats data #mcmc <- bfw(project.data = bfw::Cats, # y = "Alignment", # x = "Ratings,Reward", # saved.steps = 50000, # jags.model = "softmax", # jags.seed = 100) ## Conduct binominal generalized linear model #model <- glm(Alignment ~ Ratings + Reward, data=bfw::Cats, family = binomial(link="logit")) ## Print output from softmax #mcmc$summary.MCMC # ## Mean Median Mode ESS HDIlo HDIhi n ##beta[1,1]: Evil vs. Ratings 0.000 0.00 -0.000607 0 0.000 0.000 2000 ##beta[1,2]: Evil vs. Reward 0.000 0.00 -0.000607 0 0.000 0.000 2000 ##beta[2,1]: Good vs. Ratings 1.289 1.29 1.283403 19614 1.187 1.387 2000 ##beta[2,2]: Good vs. Reward 1.276 1.27 1.279209 20807 0.961 1.597 2000 ##beta0[1]: Intercept: Evil 0.000 0.00 -0.000607 0 0.000 0.000 2000 ##beta0[2]: Intercept: Good -7.690 -7.68 -7.659198 17693 -8.472 -6.918 2000 ##zbeta[1,1]: Evil vs. Ratings 0.000 0.00 -0.000607 0 0.000 0.000 2000 ##zbeta[1,2]: Evil vs. Reward 0.000 0.00 -0.000607 0 0.000 0.000 2000 ##zbeta[2,1]: Good vs. Ratings 2.476 2.47 2.464586 19614 2.280 2.664 2000 ##zbeta[2,2]: Good vs. Reward 0.501 0.50 0.501960 20807 0.377 0.626 2000 ##zbeta0[1]: Intercept: Evil 0.000 0.00 -0.000607 0 0.000 0.000 2000 ##zbeta0[2]: Intercept: Good -1.031 -1.03 -1.024178 22812 -1.185 -0.870 2000 # ## Print (truncated) output from GML ## Estimate Std. Error z value Pr(>|z|) ##(Intercept) -6.39328 0.27255 -23.457 < 2e-16 *** ##Ratings 1.28480 0.05136 25.014 < 2e-16 *** ##RewardAffection 1.26975 0.16381 7.751 9.1e-15 ***
The function provide a summary of each parameter of interest (mean, median, mode, effective sample size (ESS), HDI and n)
SumMCMC( par, par.names, job.names = NULL, job.group = NULL, credible.region = 0.95, ROPE = NULL, n.data, ... )SumMCMC( par, par.names, job.names = NULL, job.group = NULL, credible.region = 0.95, ROPE = NULL, n.data, ... )
par |
defined parameter |
par.names |
parameter names |
job.names |
names of all parameters in analysis, Default: NULL |
job.group |
for some hierarchical models with several layers of parameter names (e.g., latent and observed parameters), Default: NULL |
credible.region |
summarize uncertainty by defining a region of most credible values (e.g., 95 percent of the distribution), Default: 0.95 |
ROPE |
define range for region of practical equivalence (e.g., c(-0.05 , 0.05), Default: NULL |
n.data |
sample size for each parameter |
... |
further arguments passed to or from other methods |
Compute sum to zero values across all levels of a data matrix
SumToZero(q.levels, data, contrasts)SumToZero(q.levels, data, contrasts)
q.levels |
number of levels of each variable/column |
data |
data matrix to combine from |
contrasts |
specified contrasts columns |
data <- matrix(c(1,2),ncol=2) colnames(data) <- c("m1[1]","m1[2]") SumToZero( 2 , data , contrasts = NULL ) # b0[1] b1[1] b1[2] # m1[1] 1.5 -0.5 0.5data <- matrix(c(1,2),ncol=2) colnames(data) <- c("m1[1]","m1[2]") SumToZero( 2 , data , contrasts = NULL ) # b0[1] b1[1] b1[2] # m1[1] 1.5 -0.5 0.5
Small function that clears up messy code
TidyCode(tidy.code, jags = TRUE)TidyCode(tidy.code, jags = TRUE)
tidy.code |
Messy code that needs cleaning |
jags |
logical, if TRUE run code as JAGS model, Default: TRUE |
(Somewhat) tidy code
messy <- "code <- function( x ) { print (x ) }" cat(messy) code <- function( x ) { print (x ) } cat ( TidyCode(messy, jags = FALSE) ) code <- function(x) { print(x) }messy <- "code <- function( x ) { print (x ) }" cat(messy) code <- function( x ) { print (x ) } cat ( TidyCode(messy, jags = FALSE) ) code <- function(x) { print(x) }
remove excess whitespace from string
Trim(s, multi = TRUE)Trim(s, multi = TRUE)
s |
string |
multi |
logical, indicating whether or not to remove excess whitespace between characters, Default: TRUE |
Trim(" Trimmed string") # [1] "Trimmed string" Trim(" Trimmed string", FALSE) # [1] "Trimmed string"Trim(" Trimmed string") # [1] "Trimmed string" Trim(" Trimmed string", FALSE) # [1] "Trimmed string"
Extends strsplit by trimming and unlisting string
TrimSplit( x, sep = ",", fixed = FALSE, perl = FALSE, useBytes = FALSE, rm.empty = TRUE )TrimSplit( x, sep = ",", fixed = FALSE, perl = FALSE, useBytes = FALSE, rm.empty = TRUE )
x |
string |
sep |
symbol to separate data (e.g., comma-delimited), Default: ',' |
fixed |
logical, if TRUE match split exactly, otherwise use regular expressions. Has priority over perl, Default: FALSE |
perl |
logical, indicating whether or not to use Perl-compatible regexps, Default: FALSE |
useBytes |
logical, if TRUE the matching is done byte-by-byte rather than character-by-character, Default: FALSE |
rm.empty |
logical. indicating whether or not to remove empty elements, Default: TRUE |
TrimSplit("Data 1, Data2, Data3") # [1] "Data 1" "Data2" "Data3"TrimSplit("Data 1, Data2, Data3") # [1] "Data 1" "Data2" "Data3"
extending gsub by matching pattern and replacement from two vectors
VectorSub(pattern, replacement, string)VectorSub(pattern, replacement, string)
pattern |
vector containing words to match |
replacement |
vector containing words to replace existing words. |
string |
string to replace from |
modified string with replaced values
pattern <- c("A","B","C") replacement <- 1:3 string <- "A went to B went to C" VectorSub(pattern,replacement,string) # [1] "1 went to 2 went to 3"pattern <- c("A","B","C") replacement <- 1:3 string <- "A went to B went to C" VectorSub(pattern,replacement,string) # [1] "1 went to 2 went to 3"