Helpers

Description

GUI helpers

Usage

.assign.classnames()

Apply a Stratified test to a Contingency Table

Description

Applies and adds a hypothesis test to a contingency.tables object.

Usage

add.cross.strata.test(tables,name,htests,types=c("asymptotic","monte.carlo","exact"))

Arguments

Value

A contingency.tables object identical to tables, but with the test applied to each table.

See Also

add.mantel.haenszel add.test

Examples

dat<-data.frame(a=rnorm(100)>.5,b=rnorm(100)>0,c=rnorm(100)>(-.5))
tables<-contingency.tables(
 row.vars=a,
 col.vars=b,
 stratum.var=c,data=dat)
add.cross.strata.test(tables,"Mantel-Haenszel",list(function(x) mantelhaen.test(x,correct=FALSE)),
 "asymptotic")
tables

Apply the Mantel-Haenszel test to a Contingency Table

Description

Applies and adds the Cochran-Mantel-Haenzsel test to a contingency.tables object. The Cochran-Mantel-Haenzsel tests the independence of two nominal variables, stratified by a third nominal variable, assuming no three way interaction.

Usage

add.mantel.haenszel(tables,conservative=FALSE)

Arguments

Details

This is a convenience function wrapping mantelhaen.test in a add.cross.strata.test call. See mantelhaen.test for further details.

Value

A contingency.tables object identical to tables, but with the test applied to each table.

See Also

add.cross.strata.test add.test mantelhaen.test

Examples

dat<-data.frame(a=rnorm(100)>.5,b=rnorm(100)>0,c=rnorm(100)>(-.5))
tables1<-contingency.tables(
 row.vars=a,
 col.vars=b,
 stratum.var=c,data=dat)
tables1<-add.mantel.haenszel(tables1)
print(tables1,prop.r=TRUE,prop.c=TRUE,prop.t=FALSE)

Apply a test to a Contingency Tables object

Description

Applies and adds a test to a contingency.tables object.

Usage

add.test(tables,name,htests,types=c("asymptotic","monte.carlo","exact"))

add.chi.squared(tables, simulate.p.value = FALSE, B = 10000)

add.likelihood.ratio(tables, conservative = FALSE, simulate.p.value = FALSE, B = 10000)

add.fishers.exact(tables, simulate.p.value = FALSE, B = 10000)

add.correlation(tables,method=c("spearman","kendall"))

add.kruskal(tables,nominal=c("both","rows","cols"))

Arguments

Details

add.test applies a supplied list of tests to all of the tables in tables.

add.chi.squared is a wrapper function applying the chisq.test function to each table. add.likelihood.ratio is a wrapper function applying the likelihood.test function to each table. add.fishers.exact is a wrapper function applying the fisher.test function to each table. add.correlation is a wrapper function applying the cor.test function to each table. add.kruskal is a wrapper function applying the kruskal.test function to each table.

Value

A contingency.tables object identical to tables, but with the test applied to each table.

See Also

add.cross.strata.test likelihood.test cor.test kruskal.test

Examples

dat<-data.frame(a=rnorm(100)>.5,b=rnorm(100)>0,c=rnorm(100)>(-.5))
tables<-contingency.tables(
	row.vars=a,
	col.vars=b,
	stratum.var=c,data=dat)
tables<-add.chi.squared(tables,simulate.p.value=TRUE,B=10000)
tables<-add.likelihood.ratio(tables)
tables<-add.fishers.exact(tables)
tables<-add.correlation(tables,method='kendall')
tables<-add.kruskal(tables)
tables<-add.mantel.haenszel(tables)
print(tables)
remove(tables)

as.matrix method

Description

as matrix

Usage

## S3 method for class 'cor.matrix'
as.matrix(x,...)

Arguments

Value

a matrix

Non-central Chi-Squared Confidence Interval

Description

Confidence interval for the Non-centrality parameter of Non-central chi-squared distribution

Usage

chi.noncentral.conf(chival,df,conf,prec=.00001)

Arguments

Value

A 2X2 matrix whose rows represent the upper and lower bounds, and whose columns represent the parameter value and upper tail percentiles.

References

Smithson, M.J. (2003). Confidence Intervals, Quantitative Applications in the Social Sciences Series, No. 140. Thousand Oaks, CA: Sage.

See Also

Chisquare

Examples

chi.noncentral.conf(6,1,.95)
#            Result:

#      Non-Central          %
#Lower   0.2089385 0.97500899
#Upper  19.4443359 0.02499302

contin.tests.to.table

Description

Makes a nice table out of a contin.tests object

Usage

contin.tests.to.table(tests,test.digits=3,...)

Arguments

Value

A nice table

Contingency Tables

Description

Creates a contingency.tables object

Usage

contingency.tables(row.vars, col.vars, stratum.var, data=NULL, missing.include=FALSE )

Arguments

Value

A list with class "contingency.tables." Each element of the list is a single contingency table of class "contin.table" corresponding to each combination of elements of row.vars and col.vars stratified by stratum.var

See Also

extract.counts

Examples

temp.data<-data.frame(a=rnorm(100)>0,b=rnorm(100)>0,gender=rep(c("male","female"),50))
#a vs. b stratified by gender
tab<-contingency.tables(a,b,gender,data=temp.data)
tab

##add in chi-squared tests
tab<-add.chi.squared(tab)
tab

cor.matrix

Description

Creates a correlation matrix

Usage

cor.matrix(variables,with.variables,data=NULL,test=cor.test,...)

Arguments

Value

A multi.test object, representing a table of the results of func applied to each of the variables.

See Also

cor.test as.matrix.cor.matrix

Examples

dat<-data.frame(aa=rnorm(100),bb=rnorm(100),cc=rnorm(100),dd=rnorm(100))
dat$aa<-dat$aa+dat$dd
dat$cc<-dat$cc+dat$aa
cor.matrix(dat,test=cor.test)
cor.matrix(d(aa,cc),data=dat,test=cor.test,method="kendall")
cor.matrix(d(aa,cc),d(dd,bb),data=dat,test=cor.test,method="spearman")

wrapper for data.frame

Description

This function creates data frames, tightly coupled collections of variables which share many of the properties of matrices and of lists, used as the fundamental data structure by most of R's modeling software. It is a keystroke saving wrapper for the data.frame function. The only difference is that check.names and stringsAsFactors are FALSE by default.

Usage

d(..., row.names = NULL, check.rows = FALSE,
                check.names = FALSE,
                stringsAsFactors = FALSE)

Arguments

See Also

data.frame

Examples

x <- d(rnorm(10),1:10)

GUI Access functions

Description

splits a variable into two groups

Usage

deducer(cmd=NULL)

data.viewer()

Arguments

Controls Deducer's command line menus

Description

Controls Deducer's command line menus

Usage

deducer.addMenu(name, pos=length(menus)+1)
deducer.setMenus(newMenus)
deducer.getMenus()
deducer.addMenuItem(name, pos=NULL, command, menuName, silent=TRUE)
menuFunctions()

Arguments

Examples

#add a menu with two items
deducer.addMenu("TestMenu")
deducer.addMenuItem("test1",,"cat('test1 selected')","TestMenu")
deducer.addMenuItem("test2",,"print(summary(lm(rnorm(100)~rnorm(100))))","TestMenu")

#Add menu to gui if applicable
if(.windowsGUI){
	winMenuAdd("TestMenu")
	winMenuAddItem("TestMenu", "test1", "cat('test1 selected')")
	winMenuAddItem("TestMenu", "test2", "print(summary(lm(rnorm(100)~rnorm(100))))")
}else if(.jgr){
	jgr.addMenu("TestMenu")
	jgr.addMenuItem("TestMenu", "test1", "cat('test1 selected')")
	jgr.addMenuItem("TestMenu", "test2", "print(summary(lm(rnorm(100)~rnorm(100))))")	
}

Table of Descriptives

Description

Table of descriptive statistics, possibly stratified

Usage

descriptive.table(vars,
	strata,
	data,
	func.names = c("Mean","St. Deviation","Median",
		"25th Percentile","75th Percentile",
		"Minimum","Maximum","Skew","Kurtosis","Valid N"),
	func.additional)

Arguments

Value

Returns a list of matrix objects containing descriptive information on all variables in dat. One for each level or combination of levels in strata.

See Also

frequencies mean by

Examples

data(mtcars)
##means and standard deviations
descriptive.table(vars = d(mpg,hp),data= mtcars,
	func.names =c("Mean","St. Deviation","Valid N"))
##stratifying by cyl
descriptive.table(vars = d(mpg,hp) ,
	strata = d(cyl),data= mtcars,
	func.names =c("Mean","St. Deviation","Valid N"))
	
	
func.list=list(mean.deviance=function(x) mean(abs(x-mean(x))))

##Adding deviance as a statistic
descriptive.table(vars = d(mpg,hp) ,
	strata = d(cyl),data= mtcars,
	func.names =c("Mean","St. Deviation","Valid N"),func.additional=func.list)

Deducer's plug-in development tools

Description

functions pertaining to GUI development

Usage

addComponent(container, component, top, right, bottom, 
	left, topType = "REL", rightType = "REL", bottomType = "REL", 
	leftType = "REL")
getSize(component)
setSize(component,width,height)
execute(cmd)

ButtonGroupWidget 
CheckBoxesWidget 
DeducerMain 
JLabel 
RDialog
SimpleRDialog
SimpleRSubDialog 
SingleVariableWidget 
SliderWidget 
TextAreaWidget
VariableListWidget 
VariableSelectorWidget
ComboBoxWidget
RDialogMonitor
ListWidget
AddRemoveButtons
TextFieldWidget
ObjectChooserWidget

Arguments

Details

addComponent adds a Java object of class Component to a container (usually an RDialog or SimpleRDialog). the location of the component is determined by the top, right, bottom, and left arguments, which are numbers between 1 and 1000 indicating the distance from either the top (or left) of the container, with 1000 indicating the opposite side of the container. Each side can be constrained in three different ways. If the Type is "REL", the side will scale proportional to the container when the container is resized. If it is "ABS", it is not rescaled. If it is "NONE", the location of that side is determined by the componet's preferred size, which can be set with the "setPreferedSize" method.

getSize gets the height and width

setSize sets the height and width

execute executes a character representing a command, as if it were entered into the console

The rest of the items are references to the Java classes of commonly used GUI components. see www.deducer.org for more details and usage.

dich

Description

splits a variable into two groups

Usage

dich(variables,data=NULL,cut=NULL,group1=NULL,group2=NULL)

Arguments

Value

a data.frame containing the variables, recoded into two groups.

Extract Contingency Table Arrays

Description

Extracts the counts of a contingency.tables object

Usage

extract.counts(tables)

Arguments

Value

A named list of three dimensional arrays. One for each contin.table in tables

See Also

contingency.tables

Examples

temp.data<-data.frame(a=rnorm(100)>0,b=rnorm(100)>0,gender=rep(c("male","female"),50))
#a vs. b stratified by gender
tab<-contingency.tables(a,b,gender,data=temp.data)
tab

##extract counts
extract.counts(tab)

##Yields something like the following:
#$`a by b`
#, , female
#
#      FALSE TRUE
#FALSE    11    9
#TRUE     15   15
#
#, , male
#
#      FALSE TRUE
#FALSE    10   10
#TRUE     22    8

Frequency Tables

Description

Creates a set of frequency tables.

Usage

frequencies(data,r.digits=1)

Arguments

Value

Returns a list of freq.table objects. One for each variable in data.

See Also

table xtabs descriptive.table prop.table

Examples

dat<-data.frame(rnorm(100)>0,trunc(runif(100,0,5)))
##rounding to 1
frequencies(dat)
##rounding to 4
frequencies(dat,4)

get objects

Description

Enumerates all objects of a certain class

Usage

get.objects(cn,env = globalenv(),includeInherited=TRUE)

Arguments

Value

a character vector

Correlation matrix

Description

Plots a correlation matrix

Usage

ggcorplot(cor.mat,data=NULL,lines=TRUE,line.method=c("lm","loess"),type="points",
		alpha=.25,main="auto",var_text_size=5,
		cor_text_limits=c(5,25),level=.05)

Arguments

Author(s)

Mike Lawrence and Ian Fellows

See Also

cor.matrix qscatter_array

Examples

data(mtcars)
corr.mat1<-cor.matrix(variables=d(mpg,carb,carb+rnorm(length(carb))),,
	 data=mtcars,
	 test=cor.test,
	 method='spearman',
	alternative="two.sided",exact=FALSE)
	
p<-ggcorplot(corr.mat1,data = mtcars)
print(p)
## Not run: 

has.hex<-require("hexbin")
if(has.hex){
	data(diamonds)
	corr.mat<-cor.matrix(variables=d(price,carat,color),,
		 data=diamonds,
		 test=cor.test,
		 method='spearman',
		alternative="two.sided")

	p1 <- ggcorplot(cor.mat=corr.mat,data=diamonds,type="bins",
		cor_text_limits=c(5,15),
		lines=FALSE)
	print(p1)
	rm('corr.mat')

}

## End(Not run)

K Sample Test

Description

Performs a K independent sample test.

Usage

k.sample.test(formula,data,test=oneway.test,...)

Arguments

Value

A multi.test object, representing a table of the results of func applied to each of the variables.

See Also

oneway.test kruskal.test wilcox.test

Examples

dat<-data.frame(a=rnorm(100),b=rnorm(100),c=rnorm(100),d=cut(rnorm(100),4))
k.sample.test(d(a,b)~d,dat)
k.sample.test(dat[,-4]~dat$d,var.equal=TRUE)
k.sample.test(d(a,c)~d,dat,kruskal.test)

Likelihood Ratio (G test) for contingency tables

Description

Performs a likelihood ratio test of independence

Usage

likelihood.test(x,y=NULL,conservative=FALSE)

Arguments

Value

A list with class "htest" containing the following components:

Author(s)

Pete Hurd and Ian Fellows

See Also

chisq.test

Examples

data(InsectSprays)
likelihood.test(InsectSprays$count>7,InsectSprays$spray)

multi.test

Description

Creates a table from a list of htests

Usage

multi.test(tests)

Arguments

Value

A multi.test object, representing a table of the htest objects.

One Sample Test

Description

Performs a one sample test.

Usage

one.sample.test(variables,data=NULL,test=t.test,...)

Arguments

Value

A multi.test object, representing a table of the results of test applied to each of the variables.

See Also

t.test shapiro.test

Examples

data(anorexia)

#are subjects' weights at baseline and endpoint significantly different from normal
one.sample.test(variables=d(Prewt,Postwt),
	data=anorexia,
	test=shapiro.test)

#does CBT work at increasing mean wt
anorexia.sub<-subset(anorexia,Treat=="CBT")
one.sample.test(variables=Postwt-Prewt,
	data=anorexia.sub,
	test=t.test)

onesample.plot

Description

plots for one sample tests

Usage

onesample.plot(variables,data=NULL,test.value,scale=FALSE,type="hist",alpha=.2)

Arguments

Examples

data(mtcars)
onesample.plot(variables=d(mpg,cyl,disp,hp,drat,wt,qsec,vs,am,
gear,carb),data=mtcars,type='hist')

onesample.plot(variables=d(mpg,cyl,disp,hp,drat,wt,qsec,vs,am,
gear,carb),data=mtcars,type='box',alpha=1)

One Way PLot

Description

plots a categorical variable against a series of continuous variables

Usage

oneway.plot(formula,data=NULL,alpha=.2,
		box=TRUE,points=TRUE,scale=FALSE)

Arguments

Value

a ggplot object

Examples

oneway.plot(d(DriversKilled, drivers, front, rear,   kms, PetrolPrice)~law,as.data.frame(Seatbelts))

Vector Permutations

Description

Enumerates all permutations of a vector

Usage

perm(vec,duplicates=FALSE)

Arguments

Value

Returns a matrix where each row is a permutation of vec. All possible permutations are listed, and if duplicates=TRUE non-unique permutations are also listed.

See Also

sample

Examples

perm(1:4)
perm(LETTERS[4:8])

Permutation t-test

Description

Two Sample t-test via monte-carlo permutation

Usage

perm.t.test(x,y,statistic=c("t","mean"),
			alternative=c("two.sided", "less", "greater"), midp=TRUE, B=10000)

Arguments

Details

This function performs a two sample permutation test. If the mean is permuted, then the test assumes exchangability between the two samples. if the t-statistic is used, the test assumes either exchangability or a sufficiently large sample size. Because there is little lost in the way of power, and the assumptions are weaker, the t-statistic is used by default.

Value

A list with class "htest" containing the following components:

See Also

t.test

Examples

perm.t.test(rnorm(100),runif(100,-.5,.5))

Plot method

Description

Produces a circle plot for an object of class "plot.cor.matrix"

Usage

## S3 method for class 'cor.matrix'
plot(x,y=NULL,size=10,...)

Arguments

Value

a ggplot object

Print method

Description

Print object of class "contin.table" in nice layout.

Usage

## S3 method for class 'contin.table'
print(
	x,digits=3,prop.r=TRUE,prop.c=TRUE,prop.t=TRUE,
	expected.n=FALSE,residuals=FALSE,std.residuals=FALSE,
	adj.residuals=FALSE,no.tables=FALSE,...)

Arguments

Value

none

Author(s)

Ian Fellows based on the CrossTable function from the gmodels package maintained by Gregory R. Warnes

Print method

Description

Print object of class "contin.tests" in nice layout.

Usage

## S3 method for class 'contin.tests'
print(x,test.digits, ...)

Arguments

Value

none

Print method

Description

Print object of class "contingency.tables" in nice layout.

Usage

## S3 method for class 'contingency.tables'
print(x,digits=3,prop.r=TRUE,prop.c=TRUE,prop.t=TRUE,
						expected.n=FALSE,no.tables=FALSE,...)

Arguments

Value

none

Print method

Description

Print object of class "cor.matrix" in nice layout.

Usage

## S3 method for class 'cor.matrix'
print(x,digits=4,N=TRUE,CI=TRUE,stat=TRUE,p.value=TRUE,...)

Arguments

Value

none

Print method

Description

Print object of class "freq.table" in nice layout.

Usage

## S3 method for class 'freq.table'
print(x,...)

Arguments

Value

none

Print method

Description

Print object of class "multi.test" in nice layout.

Usage

## S3 method for class 'multi.test'
print(x,...)

Arguments

Value

none

qscatter_array

Description

Creates an array of scatterplots

Usage

qscatter_array(variables,with.variables,data,x.lab="",y.lab="",
				main="Correlation Array",common.scales=TRUE,alpha=.25)

Arguments

Examples

data(mtcars)
qscatter_array(d(cyl,disp,hp,drat),
	data=mtcars) + geom_smooth(method="lm")
qscatter_array(d(cyl,disp,hp,drat),d(wt,carb),data=mtcars,common.scales=FALSE)

Recode

Description

Recodes a set of variables according to a set of rules

Usage

recode.variables(data,recodes)

Arguments

Details

recodes contains a set of recoding rules separated by ";". There are three different types of recoding rules:

1. The simplest codes one value to another. If we wish to recode 1 into 2, we could use the rule "1->2;".

2. A range of values can be coded to a single value using "1:3->4;". This rule would code all values between 1 and 3 inclusive into 4. For factors, a value is between two levels if it is between them in the factor ordering. One sided ranges can be specified using the Lo and Hi key words (e.g."Lo:3->0; 4:Hi->1")

3. Default conditions can be coded using "else." For example, if we wish to recode all values >=0 to 1 and all values <0 to missing, we could use ("0:Hi->1; else->NA")

Value

returns a recoded data.frame

Author(s)

Ian Fellows adapted from code by John Fox

Examples

data<-data.frame(a=rnorm(100),b=rnorm(100),male=rnorm(100)>0)
recode.variables(data[c("a","b")] , "Lo:0 -> 0;0:Hi -> 1;")
data[c("male")] <- recode.variables(data[c("male")] , "1 -> 'Male';0 -> 'Female';else -> NA;")

ROC Plot for a logistic regression model

Description

Plots the ROC Curve

Usage

rocplot(logistic.model,diag=TRUE,pred.prob.labels=FALSE,prob.label.digits=3,AUC=TRUE)

Arguments

Value

a ggplot object

Author(s)

Ian Fellows adapted from the lroc function by Virasakdi Chongsuvivatwong

Examples

model.glm <- glm(formula=income>5930.5 ~ education + women + type,
	family=binomial(),data=Prestige,na.action=na.omit)
rocplot(model.glm)

Sort Data

Description

Sorts a data frame

Usage

## S3 method for class 'data.frame'
sort(x, decreasing, by, ...)

Arguments

Details

If by is a formula, or a character vector coerce-able into a formula, x is sorted by each element of the formula, with ties broken by subsequent elements. Elements preceded by a '-' indicate descending order, otherwise ascending order is used. Parentheses or any formula operator other than + and - are ignored, so sorting by a*b will sort based on the product of a and b.

If by is not a formula, a ~ is appended to the left hand side of the call, and coerced into a formula.

The decreasing argument is included for generic method consistency, and is not used.

Value

returns x, sorted.

Author(s)

Ian Fellows adapted from code by Ari Friedman and Kevin Wright

See Also

sort order

Examples

data(mtcars)

#sort by the number of cylenders
sort(mtcars, by= ~cyl)
sort(mtcars, by= cyl) #identical: no need for ~

#sort in descending order
sort(mtcars, by= -cyl)

#break ties with horse power
sort(mtcars,by= cyl +hp )
sort(mtcars,by= cyl -hp )

#randomly permute the data
sort(mtcars,by= rnorm(nrow(mtcars)) )

#reverse order
sort(mtcars,by= nrow(mtcars):1 )

#sort by squared deviation from mean hp
sort(mtcars,by= -(hp-mean(hp))^2 )
sort(mtcars,by= "-(hp-mean(hp))^2" ) #identical

Summary table for a linear model

Description

Computes the coefficients, std. errors, t values, and p-values for a linear model in the presence of possible heteroskedasticity.

Usage

summarylm(object,correlation=FALSE,symbolic.cor = FALSE,white.adjust=FALSE,...)

Arguments

Details

If white.adjust is false, the function returns a value identical to stats::summary.lm. Otherwise, robust summaries are computed

Value

A summary table

Examples

ctl <- c(4.17,5.58,5.18,6.11,4.50,4.61,5.17,4.53,5.33,5.14)
trt <- c(4.81,4.17,4.41,3.59,5.87,3.83,6.03,4.89,4.32,4.69)
group <- gl(2,10,20, labels=c("Ctl","Trt"))
weight <- c((ctl-mean(ctl))*10+mean(ctl), trt)
lm.D9 <- lm(weight ~ group)
summarylm(lm.D9,white.adjust=TRUE)

Table –> data.frame

Description

Creates a data.frame from a table

Usage

table.to.data(x)

Arguments

Value

A two column data.frame where each row is an observation and each column is a variable.

See Also

xtabs

Examples

tab<-matrix(c(4,5,6,9,7,3),ncol=3)
tab
table.to.data(tab)

Two Sample Test

Description

Performs a two independent sample test.

Usage

two.sample.test(formula,data=NULL,test=t.test,...)

Arguments

Value

A multi.test object, representing a table of the results of test applied to each of the variables.

See Also

t.test ks.test wilcox.test

Examples

dat<-data.frame(a=rnorm(100),b=rnorm(100),c=rnorm(100),d=rnorm(100)>(-.5))
two.sample.test(d(a,b) ~ d,dat,ks.test)
two.sample.test(a ~ dich(b,cut=0) ,dat,t.test)
two.sample.test(d(a^2,abs(b),c)~d,dat,wilcox.test)