| Title: | Estimation and Testing Classes Based on Package 'distr' |
|---|---|
| Description: | Evaluation (S4-)classes based on package distr for evaluating procedures (estimators/tests) at data/simulation in a unified way. |
| Authors: | Florian Camphausen [ctb] (contributed as student in the initial phase --2005), Matthias Kohl [aut, cph], Peter Ruckdeschel [cre, cph], Thomas Stabla [ctb] (contributed as student in the initial phase --2005) |
| Maintainer: | Peter Ruckdeschel <[email protected]> |
| License: | LGPL-3 |
| Version: | 2.8.2 |
| Built: | 2024-11-06 02:16:51 UTC |
| Source: | https://github.com/r-forge/distr |
distrTest provides
(S4-)classes for evaluating procedures (estimators/tests) at data/simulation in
a unified way based on package distr. This is achieved by means of the S4 class
Evaluation. The package is based on our packages distr and distrSim, hence uses distribution classes
and simulation classes as introduced there to capture the situation from which the simulations stem.
| Package: | distrTEst |
| Version: | 2.8.2 |
| Date: | 2024-01-30 |
| Depends: | R(>= 3.4), methods, graphics, setRNG(>= 2006.2-1), distrSim(>= 2.2) |
| Imports: | startupmsg, utils |
| Suggests: | distrEx(>= 2.2) |
| LazyLoad: | yes |
| License: | LGPL-3 |
| URL: | https://distr.r-forge.r-project.org/ |
| VCS/SVNRevision: | 1427 |
"Evaluation"
slots: [<name>(<class>)]
name(character), filename(character), call.ev(call),
Data(Dataclass), result(DataframeorNULL),
estimator(OptionalFunction)
"EvaluationList"
slots: [<name>(<class>)]
name(character), Elist(list)
Objects of class "Evaluation" are generated by a call to evaluate.
plot plot method for "Evaluation" and for "EvaluationList"
print,show print/show method for "Evaluation" and for "EvaluationList"
summary summary method for "Evaluation" and for "EvaluationList"
Data accessor method for "Evaluation", and, for "EvaluationList"
returns common Data
All slots are inspected / modified by corresponding
accessors / -replacement functions, e.g. call.ev(X)
or filename(X)<-"myevaluation.sav" for
an object of class "Evaluation".
You may suppress the start-up banner/message completely by setting options("StartupBanner"="off")
somewhere before loading this package by library or require in your R-code / R-session.
If option "StartupBanner" is not defined (default) or setting
options("StartupBanner"=NULL) or options("StartupBanner"="complete")
the complete start-up banner is displayed.
For any other value of option "StartupBanner" (i.e., not in c(NULL,"off","complete"))
only the version information is displayed.
The same can be achieved by wrapping the library or require call into
either suppressStartupMessages() or
onlytypeStartupMessages(.,atypes="version").
As for general packageStartupMessage's, you may also suppress all
the start-up banner by wrapping the library or require
call into suppressPackageStartupMessages() from
startupmsg-version 0.5 on.
Note: The first two numbers of package versions do not necessarily reflect package-individual development, but rather are chosen for the distrXXX family as a whole in order to ease updating "depends" information.
Global options controlling the plots and summaries of Evaluationlist
objects may be inspected / set by distrTEstoptions()
and getdistrTEstOption().
Thomas Stabla [email protected],
Florian Camphausen [email protected],
Peter Ruckdeschel [email protected],
Matthias Kohl [email protected]
Maintainer: Peter Ruckdeschel [email protected]
A vignette for packages distr, distrSim, distrTEst, and distrEx
is included into the mere documentation package distrDoc and may be called by
require("distrDoc");vignette("distr").
A homepage to this package is available under
https://distr.r-forge.r-project.org/
distr-package, distrSim-package,
setRNG
call.ev-methods
signature(object = "Evaluation"): returns the call which created the object
Data-methods
signature(object = "Evaluation"): returns the Data slot
signature(object = "EvaluationList"): returns the common Data slot of
the respective list elements
With distrTEstoptions and getdistrTEstOption you may inspect and change the global
variables used by package distrTEst.
distrTEstoptions(...) getdistrTEstOption(x)distrTEstoptions(...) getdistrTEstOption(x)
... |
any options can be defined, using name = value or by passing a list of such tagged values. |
x |
a character string holding an option name. |
Invoking distrTEstoptions() with no arguments returns a list with the current values of the options.
To access the value of a single option, one should use getdistrTEstOption("MaxNumberofSummarizedEvaluations"),
e.g., rather than distrTEstoptions("MaxNumberofSummarizedEvaluations") which is a list of length one.
distrTEstoptions() returns a list of the global options of distrTEst. distrTEstoptions("MaxNumberofSummarizedEvaluations") returns the global option MaxNumberofSummarizedEvaluations as a list of length 1. distrTEstoptions("MaxNumberofSummarizedEvaluations" = 3) sets the value of the global option MaxNumberofSummarizedEvaluations to 3.
getdistrTEstOption("MaxNumberofSummarizedEvaluations") the current value set for option MaxNumberofSummarizedEvaluations.
maximal number of evaluations plotted; defaults to 6
maximal number of evaluation dimensions plotted in parallel; defaults to 6
maximal number of evaluations summarized in parallel; defaults to 15
maximal number of evaluations printed in parallel; defaults to 15
Peter Ruckdeschel [email protected]
options, getOption,
distroptions, getdistrOption
distrTEstoptions() distrTEstoptions("MaxNumberofPlottedEvaluationDims") distrTEstoptions("MaxNumberofPlottedEvaluationDims" = 5) # or getdistrTEstOption("MaxNumberofPlottedEvaluationDims")distrTEstoptions() distrTEstoptions("MaxNumberofPlottedEvaluationDims") distrTEstoptions("MaxNumberofPlottedEvaluationDims" = 5) # or getdistrTEstOption("MaxNumberofPlottedEvaluationDims")
estimator-methods
signature(object = "Evaluation"): returns the estimator
evaluate-methods to produce objects of class "Evaluation"
object |
the data set / simulation on which the evaluation takes place |
estimator |
the estimation function used; should be able to deal with data in matrix
form |
resname |
(a vector of) character(s); the name for the univariate results or, in the case of multivariate results, and if the coordinates of the results have not yet been named, the basic name for them which is pasted to the coordinate number for each coordinate. |
name |
character; the name for the Evaluation object; by default the (R-)name of the Data set object. |
filename |
character; the filename for the Evaluation object (where it is to be saved to);
by default the filename of the Data set object which is concatenated with
the name of the estimator in |
besides the arguments determining the method dispatch, we have:
evaluate(object, estimator, ..., resname = "res",
name = as.character(substitute(object)),
filename = filename(object))
signature(object = "Dataclass", estimator = "function"):
creates an object of class "Evaluation", see there for further
information
signature(object = "Contsimulation", estimator = "function"):
creates an object of class "Evaluation", see there for further
information
When an estimator is used to data of the type "Dataclass" with the method evaluate, the result is an object of class "Evaluation".
Objects could be created by calls of the form new("Evaluation", Data, estimator, [result, name, filename, call.ev]).
It does not seem to be very useful to generate a new object this way, however.
It is to be preferred to use "evaluate" with a Dataclass object!
call.evObject of class "call": the call which created the object,
e.g.; “evaluate(Dataclassobject,mean)”
DataObject of class "Dataclass": the data set / simulation on which the evaluation takes place.
estimatorObject of class "OptionalFunction": estimation function used; this estimation function
should be able to deal with data in matrix form samplesize x obsDim and
should return either a univariate result or a vector (with named coordinates, if possible).
filenameObject of class "character": the filename of the evaluation;
by default the filename of the Dataclass object, which was called by evaluate
nameObject of class "character": the name of the evaluation; by default the name of the
Dataclass object, which was called by evaluate
resultObject of class "DataframeorNULL": the result of the evaluation of the estimation on data
no replacement possible
no replacement possible
replacement possible
replacement possible
no replacement possible
signature(.Object = "Evaluation"): initialize method
signature(object = "Evaluation"): returns a boxplot of the result
signature(object = "Evaluation"): returns the name of the data object, its filename, the estimator used
and the result
signature(object = "Evaluation"): saves the object in two files in the directory of R - one with
data, one without as comment file (see example)
signature(object = "Evaluation"): returns the name of the data object, its filename, the
estimator used and a statistical summary of the result
The saved "evaluation" can be loaded with the usual load-command, the saved comment with the function cload.
Thomas Stabla [email protected],
Florian Camphausen [email protected],
Peter Ruckdeschel [email protected],
Matthias Kohl [email protected]
Dataclass-class
Simulation-class
Contsimulation-class
load
cload
savedata-methods
plot-methods
simulate-methods
summary-methods
N <- Norm() # N is a standard normal distribution. C <- Cauchy() # C is a Cauchy distribution cs <- Contsimulation(filename = "csim", runs = 5, samplesize=5000, seed=setRNG(), distribution.id = N, distribution.c = C, rate = 0.1) simulate(cs) # Each of the 25000 random numbers is ideal (N-distributed) with # probability 0.9 and contaminated (C-distributed) with probability = 0.1 summary(cs) ev1 <- evaluate(cs, mean, resname="mean") # estimates the data with mean ev1 # bad results ev2 <- evaluate(cs,median, resname="median") # estimates the data with median ev2 # better results because median is robust savedata(ev1) # saves the evaluation with result as "csim.mean" and without result as # "csim.mean.comment" in the working directory # of R - "csim" is the # filename of the Contsimulation object, mean the name of the estimator rm(ev1) cload("csim.mean") # loads the evaluation without result - the object is called ev1.comment ev1.comment load("csim.mean") # loads the evaluation with result ev1 plot(ev1) # #clean up unlink("csim.mean") unlink("csim.mean.comment") # #another function to be evaluated: severalThings<- function(x) {list("mean"=mean(x),"sd"=sd(as.vector(x)), "mad"=mad(x))} ev3 <- evaluate(cs, severalThings, resname="several") plot(ev3) plot(ev3, ylim=c(0,10), col=c("blue","green", "red"))N <- Norm() # N is a standard normal distribution. C <- Cauchy() # C is a Cauchy distribution cs <- Contsimulation(filename = "csim", runs = 5, samplesize=5000, seed=setRNG(), distribution.id = N, distribution.c = C, rate = 0.1) simulate(cs) # Each of the 25000 random numbers is ideal (N-distributed) with # probability 0.9 and contaminated (C-distributed) with probability = 0.1 summary(cs) ev1 <- evaluate(cs, mean, resname="mean") # estimates the data with mean ev1 # bad results ev2 <- evaluate(cs,median, resname="median") # estimates the data with median ev2 # better results because median is robust savedata(ev1) # saves the evaluation with result as "csim.mean" and without result as # "csim.mean.comment" in the working directory # of R - "csim" is the # filename of the Contsimulation object, mean the name of the estimator rm(ev1) cload("csim.mean") # loads the evaluation without result - the object is called ev1.comment ev1.comment load("csim.mean") # loads the evaluation with result ev1 plot(ev1) # #clean up unlink("csim.mean") unlink("csim.mean.comment") # #another function to be evaluated: severalThings<- function(x) {list("mean"=mean(x),"sd"=sd(as.vector(x)), "mad"=mad(x))} ev3 <- evaluate(cs, severalThings, resname="several") plot(ev3) plot(ev3, ylim=c(0,10), col=c("blue","green", "red"))
Several objects of class "Evaluation" may be gathered in a list of class "EvaluationList", if they all have the same result-format and also share the same data-set.
Objects may be created by the generating function EvaluationList, i.e.;
EvaluationList(..., name0 = "a list of \"Evaluation\" objects"), where all arguments passed through ...
have to be objects of class "Evaluation", the corresponding result-slots have to contain
data.frames of identical dimension; the corresponding calls have to have identical object-arguments
(for the data set), and the corresponding Data-slots have to be identical.
name:Object of class "character": the name of the EvaluationList object
Elist:Object of class "list": the list of Evaluation objects
signature(object = "EvaluationList"): returns the list with the Evaluation objects
signature(object = "EvaluationList"): returns/modifies the name of the EvaluationList object
signature(object = "EvaluationList"): returns the common Data-slot
of one of the Evaluation objects
signature(object = "EvaluationList"): returns grouped boxplots of the results
signature(object = "EvaluationList"): for each list element returns
the name of the data object, its filename, the estimator used and the result
signature(object = "EvaluationList"): as print
signature(object = "EvaluationList"): returns the name of the data object, its filename, the
estimator used and a statistical summary of the result
Thomas Stabla [email protected],
Florian Camphausen [email protected],
Peter Ruckdeschel [email protected],
Matthias Kohl [email protected]
Dataclass-class
Simulation-class
Contsimulation-class
Evaluation-class
print-methods
plot-methods
simulate-methods
summary-methods
N <- Norm() # N is a standard normal distribution. C <- Cauchy() # C is a Cauchy distribution cs <- Contsimulation(filename = "csim", runs = 15, samplesize=500, seed=setRNG(), distribution.id = N, distribution.c = C, rate = 0.1) simulate(cs) # Each of the 25000 random numbers is ideal (N-distributed) with # probability 0.9 and contaminated (C-distributed) with probability = 0.1 summary(cs) ev1 <- evaluate(cs, mean) # estimates the data with mean ev1 # bad results ev2 <- evaluate(cs,median) # estimates the data with median ev2 # better results because median is robust savedata(ev1) # saves the EvaluationList with result as "csim.mean" and without result as # "csim.mean.comment" in the working directory # of R - "csim" is the # filename of the Contsimulation object, mean the name of the estimator rm(ev1) cload("csim.mean") # loads the EvaluationList without result - the object is called ev1.comment ev1.comment load("csim.mean") # loads the EvaluationList with result ev1 ElistObj <- EvaluationList(ev1,ev2,name0="myEvalList") plot(ElistObj,ylim=matrix(c(-0.5,0.5,0.5,4),nrow=2),main=c("location","scale")) plot(ElistObj,ylim=c(-0.5,0.5),main=c("location"),runs0=3:12,dims0=1,evals0=2) ElistObj summary(ElistObj) #clean up unlink("csim.mean") unlink("csim.mean.comment")N <- Norm() # N is a standard normal distribution. C <- Cauchy() # C is a Cauchy distribution cs <- Contsimulation(filename = "csim", runs = 15, samplesize=500, seed=setRNG(), distribution.id = N, distribution.c = C, rate = 0.1) simulate(cs) # Each of the 25000 random numbers is ideal (N-distributed) with # probability 0.9 and contaminated (C-distributed) with probability = 0.1 summary(cs) ev1 <- evaluate(cs, mean) # estimates the data with mean ev1 # bad results ev2 <- evaluate(cs,median) # estimates the data with median ev2 # better results because median is robust savedata(ev1) # saves the EvaluationList with result as "csim.mean" and without result as # "csim.mean.comment" in the working directory # of R - "csim" is the # filename of the Contsimulation object, mean the name of the estimator rm(ev1) cload("csim.mean") # loads the EvaluationList without result - the object is called ev1.comment ev1.comment load("csim.mean") # loads the EvaluationList with result ev1 ElistObj <- EvaluationList(ev1,ev2,name0="myEvalList") plot(ElistObj,ylim=matrix(c(-0.5,0.5,0.5,4),nrow=2),main=c("location","scale")) plot(ElistObj,ylim=c(-0.5,0.5),main=c("location"),runs0=3:12,dims0=1,evals0=2) ElistObj summary(ElistObj) #clean up unlink("csim.mean") unlink("csim.mean.comment")
filename-methods
signature(object = "Evaluation"): returns the filename of the evaluated object
name-methods
signature(object = "Evaluation"): returns the slot name of data object
signature(.Object = "Evaluation"): modifies the slot name of data object
auxiliary classes; may contain either a numeric vector or NULL [or a call / data.frame or NULL, respectively].
A virtual Class: No objects may be created from it.
No methods defined with class "numericorNULL", "CallorNULL", and "DataframeorNULL" in the signature.
From version 1.8, the result slot of an object of class evaluation is of type "DataframeorNULL"
Thomas Stabla [email protected],
Florian Camphausen [email protected],
Peter Ruckdeschel [email protected],
Matthias Kohl [email protected]
plot-methods
An S3 object of class c("plotInfo","DiagnInfo"), i.e., a list
containing the information needed to produce the
respective plot, which at a later stage could be used by different
graphic engines (like, e.g. ggplot) to produce the plot
in a different framework. A more detailed description will follow in
a subsequent version.
signature(x = "Evaluation", y="missing"): returns a boxplot of the result
signature(x = "EvaluationList", y="missing"): regroups the list according to
the different columns/coordinates of the result of the evaluation; for each such coordinate
a boxplot is generated containing possibly several procedures and if evaluated at a Contsimulation
object also grouped into evaluations on ideal and real data. The plot-arguments main
and ylim may be transmitted coordinatewise
(i.e.; a vector of length (result-dimension), respectively a 2 x (result-dimension) matrix)
or globally, using the usual recycling rules.
print/show-methods
signature(x = "Evaluation"): returns the name of the data object, its filename, the estimator used and the result;
optional arguments:
the indices of runs to be summarized;
the indices of result dimensions to be summarized;
internal argument:
decides if name of Dataobject and Datafile are printed out
(which is done if inLIST==FALSE); defaults to FALSE but is
TRUE when summary is called from summary-method for signature(object = "EvaluationList")
signature(object = "EvaluationList"): after printing the name of the list,
for each member of the list print is executed;
optional arguments:
the indices of evaluations to be summarized;— of this vector eval0
maximally MaxNumberofSummarizedEvaluations evaluations are summarized where MaxNumberofPrintedEvaluations
is a global option, see distrTEstoptions
the indices of runs to be summarized;
the indices of observation dimensions to be summarized;
signature(x = "Evaluation"): the same as print (without optional arguments)
signature(x = "EvaluationList"): the same as print (without optional arguments)
result-methods
signature(object = "Evaluation"): returns the result of an evaluation
savedata-methods
signature(object = "Evaluation"): saves the object in two files in the directory of R - one with
data —filename = <filename>—, one without as comment file —filename = <filename>.comment—(see example);
<filename>the filename can be specified in the optional argument fileN; by default it is concatenated
from the filename of the Dataclass object and the estimatorname, which you may either pass as argument
estimatorName or it is taken as the R-name of the corresponding R-function.
For an example, see Simulation-class
and Contsimulation-class
Dataclass-class
Simulation-class
Contsimulation-class
Evaluation-class
summary-methods
signature(object = "Evaluation"): returns the name of the data object, its filename, the estimator used and
a statistical summary of the result;optional arguments:
the indices of runs to be summarized;
the indices of result dimensions to be summarized;
internal argument:
decides if name of Dataobject and Datafile are printed out
(which is done if inLIST==FALSE); defaults to FALSE but is
TRUE when summary is called from summary-method for signature(object = "EvaluationList")
signature(object = "EvaluationList"): returns, for each member of the list a summary of
the corresponding Evaluation object;
optional arguments:
the indices of evaluations to be summarized;— of this vector eval0
maximally MaxNumberofSummarizedEvaluations evaluations are summarized where MaxNumberofSummarizedEvaluations
is a global option, see distrTEstoptions
the indices of runs to be summarized;
the indices of observation dimensions to be summarized;