R/pguModel.R
In SMLMS/pguIMP: 'pguIMP'
#' @title pgu.model
#'
#' @description
#' Comprises a list of models for data manipulation.
#'
#' @details
#' Comprises a list of pgu.normDist objects and model parameters.
#' These can be used to scale data.
#' This object is used by the shiny based gui and is not for use in individual R-scripts!
#'
#' @format [R6::R6Class] object.
#'
#' @importFrom dplyr select select_if slice
#' @importFrom ggplot2 coord_flip ggplot layer_scales scale_x_continuous
#' @importFrom ggplot2 scale_y_continuous
#' @importFrom gridExtra grid.arrange
#' @importFrom magrittr %>%
#' @importFrom R6 R6Class
#' @importFrom shiny Progress
#' @importFrom tibble tibble
#'
#' @include pguNormDist.R
#'
#' @author Sebastian Malkusch, \email{malkusch@@med.uni-frankfurt.de}
#'
#' @export
#'
pgu.model <- R6::R6Class("pgu.model",
####################
# instance variables
####################
private = list(
.modelList = "pgu.normDist",
.modelParameter = "tbl_df"
),
##################
# accessor methods
##################
active = list(
#' @field modelList
#' Returns a vector of pgu-normDist objects.
#' (pgu.normDist)
modelList = function(){
return(private$.modelList)
},
#' @field modelParameter
#' Returns a dataframe comrising model parameters.
#' (tibble::tibble)
modelParameter = function(){
return(private$.modelParameter)
}
),
###################
# memory management
###################
public = list(
#' @description
#' Creates and returns a new `pgu.model` object.
#' @param data
#' The data to be analyzed.
#' (tibble::tibble)
#' @return
#' A new `pgu.model` object.
#' (pguIMP::pgu.model)
initialize = function(data = "tbl_df"){
if(class(data)[1] != "tbl_df"){
data <- tibble::tibble(names <- "none",
values <- c(NA))
}
self$resetModel(data)
}, #function
#' @description
#' Clears the heap and
#' indicates that instance of `pgu.model` is removed from heap.
finalize = function(){
print("Instance of pgu.model removed from heap")
}, #function
############################
# print instance variables #
############################
#' @description
#' Prints instance variables of a `pgu.model` object.
#' @return
#' string
print = function(){
rString <- sprintf("\npgu.model\n")
cat(rString)
cat("modelParameter:\n")
print(self$modelParameter)
cat("\n\n")
invisible(self)
}, #function
####################
# public functions #
####################
#' @description
#' Resets instance variable `modelParameter`
#' @param data
#' Dataframe to be analyzed.
#' (tibble::tibble)
resetModelParameter = function(data = "tbl_df"){
features <- data %>%
dplyr::select_if(is.numeric) %>%
colnames()
mu <- c(rep(0.0, length(features)))
sigma <- c(rep(0.0, length(features)))
# fit quality
dataPoints <- c(rep(0.0, length(features)))
logLikelihood <- c(rep(0.0, length(features)))
degOfFreedom <- c(rep(0.0, length(features)))
bic <- c(rep(0.0, length(features)))
aic <- c(rep(0.0, length(features)))
aicc <- c(rep(0.0, length(features)))
rmse <- c(rep(0.0, length(features)))
# shapiro wilk test parameters
w.shapiro <- c(rep(0.0, length(features)))
p.shapiro <- c(rep(0.0, length(features)))
# kolmogorow smirnow test parameters
d.kolmogorow <- c(rep(0.0, length(features)))
p.kolmogorow <- c(rep(0.0, length(features)))
# anderson darling test parameters
a.anderson <- c(rep(0.0, length(features)))
p.anderson <- c(rep(0.0, length(features)))
# create modelParameter
private$.modelParameter <- tibble::tibble(features, mu, sigma, dataPoints, logLikelihood, degOfFreedom, bic, aic, aicc, rmse,
w.shapiro, p.shapiro, d.kolmogorow, p.kolmogorow, a.anderson, p.anderson)
}, #function
#' @description
#' Resets instance variable `modelList`
#' @param data
#' Dataframe to be analyzed.
#' (tibble::tibble)
resetModelList = function(data = "tbl_df"){
results=list()
for (feature in self$modelParameter[["features"]]){
idx <- self$featureIdx(feature)
if(!is.na(idx)){
results[feature] <- list(pgu.normDist$new(data[feature]))
}#if
}#for
private$.modelList <- results
}, #function
#' @description
#' Resets instance variable `modelList`.
#' Resets instance variable `modelParameter`.
#' Displays progress if shiny is loaded.
#' @param data
#' Dataframe to be analyzed.
#' (tibble::tibble)
#' @param progress
#' If shiny is loaded, the analysis' progress is stored in this instance of the shiny Progress class.
#' (shiny::Progress)
resetModel = function(data = "tbl_df", progress = "Progress"){
self$resetModelParameter(data)
self$resetModelList(data)
self$fitData(progress)
}, #function
#' @description
#' Stores the information of a pgu.norDist object in an
#' entry of the instance variable `modelList`
#' @param data
#' Instance of pgu.normDist
#' (pguIMP::pgu.normDist)
#' @param feature
#' Attribute corresponding to the pgu.normDist object data.
#' (character)
setNormDist = function(data = "pgu.normDist", feature = "character"){
idx <- match(feature, self$modelParameter[["features"]])
if(!is.na(idx)){
private$.modelList[feature] <- list(data)
self$logFitResultsFeature(feature)
}#if
}, #function
####################
# helper functions #
####################
#' @description
#' Returns the index of a pgu.normDist object wihtin the instance variable `modelParameter`.
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' Index of attribute entry in dataframe
#' (numeric)
featureIdx = function(feature = "character"){
idx <- match(feature, self$modelParameter[["features"]])
if(is.na(idx)){
rString <- sprintf("\nWarning in pgu.model: feature %s is not known\n",
feature)
cat(rString)
}#if
return(idx)
}, #function
#################
# fit functions #
#################
#' @description
#' Runs the fit function of a pgu.normDist object at a
#' user denied position within the instance variable modelList.
#' @param feature
#' Attribute's name.
#' (character)
fitFeature = function(feature = "character"){
idx <- match(feature, self$modelParameter[["features"]])
if(!is.na(idx)){
tryCatch({
private$.modelList[[feature]]$fit()
self$logFitResultsFeature(feature)
},
error = function(e){
self$logFailedFitResultsFeature(feature)
print(e)
}
)#tryCatch
}#if
}, #function
#' @description
#' Loops through all attributes and calls the object's
#' ftiFeature function.
#' Displays progress if shiny is loaded.
#' @param progress
#' If shiny is loaded, the analysis' progress is stored in this instance of the shiny Progress class.
#' (shiny::Progress)
fitData = function(progress = "Progress"){
for (feature in self$modelParameter[["features"]]){
if(("shiny" %in% (.packages())) & (class(progress)[1] == "Progress")){
progress$inc(1)
}#if
idx <- self$featureIdx(feature)
if(!is.na(idx)){
self$fitFeature(feature)
}#if
}#for
}, #function
###############
# log results #
###############
#' @description
#' Stores results from fitting procedure of a user defined attribute
#' into the corrsponding attribute of instance variable `modelParameter`.
#' @param feature
#' Attribute's name.
#' (character)
logFitResultsFeature = function(feature = "character"){
idx <- self$featureIdx(feature)
if(!is.na(idx)){
private$.modelParameter[idx, "mu"] <- private$.modelList[[feature]]$expMu
private$.modelParameter[idx, "sigma"] <- private$.modelList[[feature]]$expSigma
private$.modelParameter[idx, "logLikelihood"] <- private$.modelList[[feature]]$logLikelihood
private$.modelParameter[idx, "dataPoints"] <- private$.modelList[[feature]]$dataPoints
private$.modelParameter[idx, "degOfFreedom"] <- private$.modelList[[feature]]$degOfFreedom
private$.modelParameter[idx, "bic"] <- private$.modelList[[feature]]$bic
private$.modelParameter[idx, "aic"] <- private$.modelList[[feature]]$aic
private$.modelParameter[idx, "aicc"] <- private$.modelList[[feature]]$aicc
private$.modelParameter[idx, "rmse"] <- private$.modelList[[feature]]$rmse
private$.modelParameter[idx, "w.shapiro"] <- private$.modelList[[feature]]$w.shapiro
private$.modelParameter[idx, "p.shapiro"] <- private$.modelList[[feature]]$p.shapiro
private$.modelParameter[idx, "d.kolmogorow"] <- private$.modelList[[feature]]$d.kolmogorow
private$.modelParameter[idx, "p.kolmogorow"] <- private$.modelList[[feature]]$p.kolmogorow
private$.modelParameter[idx, "a.anderson"] <- private$.modelList[[feature]]$a.anderson
private$.modelParameter[idx, "p.anderson"] <- private$.modelList[[feature]]$p.anderson
}#if
}, #feature
#' @description
#' Stores results from fitting procedure of a user defined attribute
#' into the corrsponding attribute of instance variable `modelParameter`
#' in case of a failed fitting routine.
#' @param feature
#' Attribute's name.
#' (character)
logFailedFitResultsFeature = function(feature = "character"){
idx <- self$featureIdx(feature)
if(!is.na(idx)){
private$.modelParameter[idx, "mu"] <- NA
private$.modelParameter[idx, "sigma"] <- NA
private$.modelParameter[idx, "logLikelihood"] <- NA
private$.modelParameter[idx, "dataPoints"] <- NA
private$.modelParameter[idx, "degOfFreedom"] <- NA
private$.modelParameter[idx, "bic"] <- NA
private$.modelParameter[idx, "aic"] <- NA
private$.modelParameter[idx, "aicc"] <- NA
private$.modelParameter[idx, "rmse"] <- NA
private$.modelParameter[idx, "w.shapiro"] <- NA
private$.modelParameter[idx, "p.shapiro"] <- NA
private$.modelParameter[idx, "d.kolmogorow"] <- NA
private$.modelParameter[idx, "p.kolmogorow"] <- NA
private$.modelParameter[idx, "a.anderson"] <- NA
private$.modelParameter[idx, "p.anderson"] <- NA
}#if
}, #function
##############
# scale data #
##############
#' @description
#' Scales numeric data based upon the model of a user defined attribute.
#' @param value
#' Numeric vector
#' (numeric)
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' scaled version of the given vector
#' (numeric)
scaleNumeric = function(value = "numeric", feature = "character"){
idx <- self$featureIdx(feature)
if(!is.na(idx)){
mu <- private$.modelParameter[[idx, "mu"]]
sigma <- private$.modelParameter[[idx, "sigma"]]
value <- (value - mu)/sigma
}#if
return(value)
}, #feature
#' @description
#' Scales a dataframe based upon a list of models
#' stored in the instance variable modelList..
#' @param data
#' Dataframe to be analyzed.
#' (tibble::tibble)
#' @return
#' scaled version of the given dataframe
#' (tibble::tibble)
scaleData = function(data = "tbl_df"){
for (feature in self$modelParameter[["features"]]){
idx <- self$featureIdx(feature)
if(!is.na(idx)){
data[feature] <- self$scaleNumeric(data[[feature]], feature)
}#if
}#for
return(data)
}, #function
################
# rescale data #
################
#' @description
#' Re-scales numeric data based upon the model of a user defined attribute.
#' @param value
#' Numeric vector
#' (numeric)
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' Re-scaled version of the given vector
#' (numeric)
rescaleNumeric = function(value = "numeric", feature = "character"){
idx <- self$featureIdx(feature)
if(!is.na(idx)){
mu <- private$.modelParameter[[idx, "mu"]]
sigma <- private$.modelParameter[[idx, "sigma"]]
value <- (value * sigma) + mu
}#if
return(value)
}, #function
#' @description
#' Re-scales a dataframe based upon a list of models
#' stored in the instance variable modelList..
#' @param data
#' Dataframe to be analyzed.
#' (tibble::tibble)
#' @return
#' Re-scaled version of the given dataframe
#' (tibble::tibble)
rescaleData = function(data = "tbl_df"){
for (feature in self$modelParameter[["features"]]){
idx <- self$featureIdx(feature)
if(!is.na(idx)){
data[feature] <- self$rescaleNumeric(data[[feature]], feature)
}#if
}#for
return(data)
}, #function
##################
# return results #
##################
#' @description
#' Returns the model parameter (expectation value, standard deviation).
#' @return
#' Dataframe comprising model parameter.
#' (tibble::tibble)
modelParameterData = function(){
self$modelParameter %>%
dplyr::select(features, mu:sigma) %>%
return()
}, #function
#' @description
#' Returns the model parameter (expectation value, standard deviation)
#' for a user deined attribute.
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' Dataframe comprising model parameter.
#' (tibble::tibble)
modelParameterFeature = function(feature = "character"){
idx <- self$featureIdx(feature)
self$modelParameter %>%
dplyr::slice(idx) %>%
dplyr::select(features, mu:sigma) %>%
return()
}, #function
#' @description
#' Returns the model parameters connected to model quality.
#' @return
#' Dataframe comprising model parameter.
#' (tibble::tibble)
modelQualityData = function(){
self$modelParameter %>%
dplyr::select(features, dataPoints:rmse) %>%
return()
}, #function
#' @description
#' Returns the model parameters connected to model quality
#' for a user deined attribute.
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' Dataframe comprising model parameter.
#' (tibble::tibble)
modelQualityFeature = function(feature = "character"){
idx <- self$featureIdx(feature)
self$modelParameter %>%
dplyr::slice(idx) %>%
dplyr::select(features, dataPoints:rmse) %>%
return()
}, #function
#' @description
#' Returns the model fit results.
#' @return
#' Dataframe comprising model fit results.
#' (tibble::tibble)
fitResultData = function(){
self$modelParameter %>%
dplyr::select(features, mu:rmse) %>%
return()
}, #function
#' @description
#' Returns the model fit results
#' for a user deined attribute.
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' Dataframe comprising model fit results.
#' (tibble::tibble)
fitResultFeature = function(feature = "character"){
idx <- self$featureIdx(feature)
self$modelParameter %>%
dplyr::slice(idx) %>%
dplyr::select(features, mu:rmse) %>%
return()
}, #function
#' @description
#' Returns the hypothesis test results.
#' @return
#' Dataframe comprising the hypothesis test results.
#' (tibble::tibble)
testResultData = function(){
self$modelParameter %>%
dplyr::select(features, w.shapiro:p.anderson) %>%
return()
}, #function
#' @description
#' Returns the hypothesis test results.
#' for a user deined attribute.
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' Dataframe comprising the hypothesis test results.
#' (tibble::tibble)
testResultFeature = function(feature = "character"){
idx <- self$featureIdx(feature)
self$modelParameter %>%
dplyr::slice(idx) %>%
dplyr::select(features, w.shapiro:p.anderson) %>%
return()
}, #function
################
# plot results #
################
#' @description
#' Creates and returns a composite graphical analysis
#' of the modeling procedure of
#' a user defined attribute.
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' Composite result plot.
#' (ggplot2::ggplot)
plotModel = function(feature = "character"){
idx <- self$featureIdx(feature)
model <- self$modelList[[idx]]
p1 <- model$plotHistogram()
xLimits <- ggplot2::layer_scales(p1)$x$range$range
p2 <- model$plotResiduals() +
ggplot2::scale_x_continuous(position = "bottom", limits=xLimits)
p3 <- model$plotResidualDist() +
ggplot2::coord_flip()
yLimits <- ggplot2::layer_scales(p3)$x$range$range
p2 <- p2 +
ggplot2::scale_y_continuous(limits=yLimits)
p4 <- model$normalQQPlot()
p <- gridExtra::grid.arrange(p1,p2,p3,p4, layout_matrix = rbind(c(1,1,4),c(2,2,3))) %>%
suppressMessages()
return(p)
}#function
)#public
)#class
SMLMS/pguIMP documentation built on Oct. 1, 2021, 12:01 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
#' @title pgu.model
#'
#' @description
#' Comprises a list of models for data manipulation.
#'
#' @details
#' Comprises a list of pgu.normDist objects and model parameters.
#' These can be used to scale data.
#' This object is used by the shiny based gui and is not for use in individual R-scripts!
#'
#' @format [R6::R6Class] object.
#'
#' @importFrom dplyr select select_if slice
#' @importFrom ggplot2 coord_flip ggplot layer_scales scale_x_continuous
#' @importFrom ggplot2 scale_y_continuous
#' @importFrom gridExtra grid.arrange
#' @importFrom magrittr %>%
#' @importFrom R6 R6Class
#' @importFrom shiny Progress
#' @importFrom tibble tibble
#'
#' @include pguNormDist.R
#'
#' @author Sebastian Malkusch, \email{malkusch@@med.uni-frankfurt.de}
#'
#' @export
#'
pgu.model <- R6::R6Class("pgu.model",
####################
# instance variables
####################
private = list(
.modelList = "pgu.normDist",
.modelParameter = "tbl_df"
),
##################
# accessor methods
##################
active = list(
#' @field modelList
#' Returns a vector of pgu-normDist objects.
#' (pgu.normDist)
modelList = function(){
return(private$.modelList)
},
#' @field modelParameter
#' Returns a dataframe comrising model parameters.
#' (tibble::tibble)
modelParameter = function(){
return(private$.modelParameter)
}
),
###################
# memory management
###################
public = list(
#' @description
#' Creates and returns a new `pgu.model` object.
#' @param data
#' The data to be analyzed.
#' (tibble::tibble)
#' @return
#' A new `pgu.model` object.
#' (pguIMP::pgu.model)
initialize = function(data = "tbl_df"){
if(class(data)[1] != "tbl_df"){
data <- tibble::tibble(names <- "none",
values <- c(NA))
}
self$resetModel(data)
}, #function
#' @description
#' Clears the heap and
#' indicates that instance of `pgu.model` is removed from heap.
finalize = function(){
print("Instance of pgu.model removed from heap")
}, #function
############################
# print instance variables #
############################
#' @description
#' Prints instance variables of a `pgu.model` object.
#' @return
#' string
print = function(){
rString <- sprintf("\npgu.model\n")
cat(rString)
cat("modelParameter:\n")
print(self$modelParameter)
cat("\n\n")
invisible(self)
}, #function
####################
# public functions #
####################
#' @description
#' Resets instance variable `modelParameter`
#' @param data
#' Dataframe to be analyzed.
#' (tibble::tibble)
resetModelParameter = function(data = "tbl_df"){
features <- data %>%
dplyr::select_if(is.numeric) %>%
colnames()
mu <- c(rep(0.0, length(features)))
sigma <- c(rep(0.0, length(features)))
# fit quality
dataPoints <- c(rep(0.0, length(features)))
logLikelihood <- c(rep(0.0, length(features)))
degOfFreedom <- c(rep(0.0, length(features)))
bic <- c(rep(0.0, length(features)))
aic <- c(rep(0.0, length(features)))
aicc <- c(rep(0.0, length(features)))
rmse <- c(rep(0.0, length(features)))
# shapiro wilk test parameters
w.shapiro <- c(rep(0.0, length(features)))
p.shapiro <- c(rep(0.0, length(features)))
# kolmogorow smirnow test parameters
d.kolmogorow <- c(rep(0.0, length(features)))
p.kolmogorow <- c(rep(0.0, length(features)))
# anderson darling test parameters
a.anderson <- c(rep(0.0, length(features)))
p.anderson <- c(rep(0.0, length(features)))
# create modelParameter
private$.modelParameter <- tibble::tibble(features, mu, sigma, dataPoints, logLikelihood, degOfFreedom, bic, aic, aicc, rmse,
w.shapiro, p.shapiro, d.kolmogorow, p.kolmogorow, a.anderson, p.anderson)
}, #function
#' @description
#' Resets instance variable `modelList`
#' @param data
#' Dataframe to be analyzed.
#' (tibble::tibble)
resetModelList = function(data = "tbl_df"){
results=list()
for (feature in self$modelParameter[["features"]]){
idx <- self$featureIdx(feature)
if(!is.na(idx)){
results[feature] <- list(pgu.normDist$new(data[feature]))
}#if
}#for
private$.modelList <- results
}, #function
#' @description
#' Resets instance variable `modelList`.
#' Resets instance variable `modelParameter`.
#' Displays progress if shiny is loaded.
#' @param data
#' Dataframe to be analyzed.
#' (tibble::tibble)
#' @param progress
#' If shiny is loaded, the analysis' progress is stored in this instance of the shiny Progress class.
#' (shiny::Progress)
resetModel = function(data = "tbl_df", progress = "Progress"){
self$resetModelParameter(data)
self$resetModelList(data)
self$fitData(progress)
}, #function
#' @description
#' Stores the information of a pgu.norDist object in an
#' entry of the instance variable `modelList`
#' @param data
#' Instance of pgu.normDist
#' (pguIMP::pgu.normDist)
#' @param feature
#' Attribute corresponding to the pgu.normDist object data.
#' (character)
setNormDist = function(data = "pgu.normDist", feature = "character"){
idx <- match(feature, self$modelParameter[["features"]])
if(!is.na(idx)){
private$.modelList[feature] <- list(data)
self$logFitResultsFeature(feature)
}#if
}, #function
####################
# helper functions #
####################
#' @description
#' Returns the index of a pgu.normDist object wihtin the instance variable `modelParameter`.
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' Index of attribute entry in dataframe
#' (numeric)
featureIdx = function(feature = "character"){
idx <- match(feature, self$modelParameter[["features"]])
if(is.na(idx)){
rString <- sprintf("\nWarning in pgu.model: feature %s is not known\n",
feature)
cat(rString)
}#if
return(idx)
}, #function
#################
# fit functions #
#################
#' @description
#' Runs the fit function of a pgu.normDist object at a
#' user denied position within the instance variable modelList.
#' @param feature
#' Attribute's name.
#' (character)
fitFeature = function(feature = "character"){
idx <- match(feature, self$modelParameter[["features"]])
if(!is.na(idx)){
tryCatch({
private$.modelList[[feature]]$fit()
self$logFitResultsFeature(feature)
},
error = function(e){
self$logFailedFitResultsFeature(feature)
print(e)
}
)#tryCatch
}#if
}, #function
#' @description
#' Loops through all attributes and calls the object's
#' ftiFeature function.
#' Displays progress if shiny is loaded.
#' @param progress
#' If shiny is loaded, the analysis' progress is stored in this instance of the shiny Progress class.
#' (shiny::Progress)
fitData = function(progress = "Progress"){
for (feature in self$modelParameter[["features"]]){
if(("shiny" %in% (.packages())) & (class(progress)[1] == "Progress")){
progress$inc(1)
}#if
idx <- self$featureIdx(feature)
if(!is.na(idx)){
self$fitFeature(feature)
}#if
}#for
}, #function
###############
# log results #
###############
#' @description
#' Stores results from fitting procedure of a user defined attribute
#' into the corrsponding attribute of instance variable `modelParameter`.
#' @param feature
#' Attribute's name.
#' (character)
logFitResultsFeature = function(feature = "character"){
idx <- self$featureIdx(feature)
if(!is.na(idx)){
private$.modelParameter[idx, "mu"] <- private$.modelList[[feature]]$expMu
private$.modelParameter[idx, "sigma"] <- private$.modelList[[feature]]$expSigma
private$.modelParameter[idx, "logLikelihood"] <- private$.modelList[[feature]]$logLikelihood
private$.modelParameter[idx, "dataPoints"] <- private$.modelList[[feature]]$dataPoints
private$.modelParameter[idx, "degOfFreedom"] <- private$.modelList[[feature]]$degOfFreedom
private$.modelParameter[idx, "bic"] <- private$.modelList[[feature]]$bic
private$.modelParameter[idx, "aic"] <- private$.modelList[[feature]]$aic
private$.modelParameter[idx, "aicc"] <- private$.modelList[[feature]]$aicc
private$.modelParameter[idx, "rmse"] <- private$.modelList[[feature]]$rmse
private$.modelParameter[idx, "w.shapiro"] <- private$.modelList[[feature]]$w.shapiro
private$.modelParameter[idx, "p.shapiro"] <- private$.modelList[[feature]]$p.shapiro
private$.modelParameter[idx, "d.kolmogorow"] <- private$.modelList[[feature]]$d.kolmogorow
private$.modelParameter[idx, "p.kolmogorow"] <- private$.modelList[[feature]]$p.kolmogorow
private$.modelParameter[idx, "a.anderson"] <- private$.modelList[[feature]]$a.anderson
private$.modelParameter[idx, "p.anderson"] <- private$.modelList[[feature]]$p.anderson
}#if
}, #feature
#' @description
#' Stores results from fitting procedure of a user defined attribute
#' into the corrsponding attribute of instance variable `modelParameter`
#' in case of a failed fitting routine.
#' @param feature
#' Attribute's name.
#' (character)
logFailedFitResultsFeature = function(feature = "character"){
idx <- self$featureIdx(feature)
if(!is.na(idx)){
private$.modelParameter[idx, "mu"] <- NA
private$.modelParameter[idx, "sigma"] <- NA
private$.modelParameter[idx, "logLikelihood"] <- NA
private$.modelParameter[idx, "dataPoints"] <- NA
private$.modelParameter[idx, "degOfFreedom"] <- NA
private$.modelParameter[idx, "bic"] <- NA
private$.modelParameter[idx, "aic"] <- NA
private$.modelParameter[idx, "aicc"] <- NA
private$.modelParameter[idx, "rmse"] <- NA
private$.modelParameter[idx, "w.shapiro"] <- NA
private$.modelParameter[idx, "p.shapiro"] <- NA
private$.modelParameter[idx, "d.kolmogorow"] <- NA
private$.modelParameter[idx, "p.kolmogorow"] <- NA
private$.modelParameter[idx, "a.anderson"] <- NA
private$.modelParameter[idx, "p.anderson"] <- NA
}#if
}, #function
##############
# scale data #
##############
#' @description
#' Scales numeric data based upon the model of a user defined attribute.
#' @param value
#' Numeric vector
#' (numeric)
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' scaled version of the given vector
#' (numeric)
scaleNumeric = function(value = "numeric", feature = "character"){
idx <- self$featureIdx(feature)
if(!is.na(idx)){
mu <- private$.modelParameter[[idx, "mu"]]
sigma <- private$.modelParameter[[idx, "sigma"]]
value <- (value - mu)/sigma
}#if
return(value)
}, #feature
#' @description
#' Scales a dataframe based upon a list of models
#' stored in the instance variable modelList..
#' @param data
#' Dataframe to be analyzed.
#' (tibble::tibble)
#' @return
#' scaled version of the given dataframe
#' (tibble::tibble)
scaleData = function(data = "tbl_df"){
for (feature in self$modelParameter[["features"]]){
idx <- self$featureIdx(feature)
if(!is.na(idx)){
data[feature] <- self$scaleNumeric(data[[feature]], feature)
}#if
}#for
return(data)
}, #function
################
# rescale data #
################
#' @description
#' Re-scales numeric data based upon the model of a user defined attribute.
#' @param value
#' Numeric vector
#' (numeric)
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' Re-scaled version of the given vector
#' (numeric)
rescaleNumeric = function(value = "numeric", feature = "character"){
idx <- self$featureIdx(feature)
if(!is.na(idx)){
mu <- private$.modelParameter[[idx, "mu"]]
sigma <- private$.modelParameter[[idx, "sigma"]]
value <- (value * sigma) + mu
}#if
return(value)
}, #function
#' @description
#' Re-scales a dataframe based upon a list of models
#' stored in the instance variable modelList..
#' @param data
#' Dataframe to be analyzed.
#' (tibble::tibble)
#' @return
#' Re-scaled version of the given dataframe
#' (tibble::tibble)
rescaleData = function(data = "tbl_df"){
for (feature in self$modelParameter[["features"]]){
idx <- self$featureIdx(feature)
if(!is.na(idx)){
data[feature] <- self$rescaleNumeric(data[[feature]], feature)
}#if
}#for
return(data)
}, #function
##################
# return results #
##################
#' @description
#' Returns the model parameter (expectation value, standard deviation).
#' @return
#' Dataframe comprising model parameter.
#' (tibble::tibble)
modelParameterData = function(){
self$modelParameter %>%
dplyr::select(features, mu:sigma) %>%
return()
}, #function
#' @description
#' Returns the model parameter (expectation value, standard deviation)
#' for a user deined attribute.
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' Dataframe comprising model parameter.
#' (tibble::tibble)
modelParameterFeature = function(feature = "character"){
idx <- self$featureIdx(feature)
self$modelParameter %>%
dplyr::slice(idx) %>%
dplyr::select(features, mu:sigma) %>%
return()
}, #function
#' @description
#' Returns the model parameters connected to model quality.
#' @return
#' Dataframe comprising model parameter.
#' (tibble::tibble)
modelQualityData = function(){
self$modelParameter %>%
dplyr::select(features, dataPoints:rmse) %>%
return()
}, #function
#' @description
#' Returns the model parameters connected to model quality
#' for a user deined attribute.
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' Dataframe comprising model parameter.
#' (tibble::tibble)
modelQualityFeature = function(feature = "character"){
idx <- self$featureIdx(feature)
self$modelParameter %>%
dplyr::slice(idx) %>%
dplyr::select(features, dataPoints:rmse) %>%
return()
}, #function
#' @description
#' Returns the model fit results.
#' @return
#' Dataframe comprising model fit results.
#' (tibble::tibble)
fitResultData = function(){
self$modelParameter %>%
dplyr::select(features, mu:rmse) %>%
return()
}, #function
#' @description
#' Returns the model fit results
#' for a user deined attribute.
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' Dataframe comprising model fit results.
#' (tibble::tibble)
fitResultFeature = function(feature = "character"){
idx <- self$featureIdx(feature)
self$modelParameter %>%
dplyr::slice(idx) %>%
dplyr::select(features, mu:rmse) %>%
return()
}, #function
#' @description
#' Returns the hypothesis test results.
#' @return
#' Dataframe comprising the hypothesis test results.
#' (tibble::tibble)
testResultData = function(){
self$modelParameter %>%
dplyr::select(features, w.shapiro:p.anderson) %>%
return()
}, #function
#' @description
#' Returns the hypothesis test results.
#' for a user deined attribute.
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' Dataframe comprising the hypothesis test results.
#' (tibble::tibble)
testResultFeature = function(feature = "character"){
idx <- self$featureIdx(feature)
self$modelParameter %>%
dplyr::slice(idx) %>%
dplyr::select(features, w.shapiro:p.anderson) %>%
return()
}, #function
################
# plot results #
################
#' @description
#' Creates and returns a composite graphical analysis
#' of the modeling procedure of
#' a user defined attribute.
#' @param feature
#' Attribute's name.
#' (character)
#' @return
#' Composite result plot.
#' (ggplot2::ggplot)
plotModel = function(feature = "character"){
idx <- self$featureIdx(feature)
model <- self$modelList[[idx]]
p1 <- model$plotHistogram()
xLimits <- ggplot2::layer_scales(p1)$x$range$range
p2 <- model$plotResiduals() +
ggplot2::scale_x_continuous(position = "bottom", limits=xLimits)
p3 <- model$plotResidualDist() +
ggplot2::coord_flip()
yLimits <- ggplot2::layer_scales(p3)$x$range$range
p2 <- p2 +
ggplot2::scale_y_continuous(limits=yLimits)
p4 <- model$normalQQPlot()
p <- gridExtra::grid.arrange(p1,p2,p3,p4, layout_matrix = rbind(c(1,1,4),c(2,2,3))) %>%
suppressMessages()
return(p)
}#function
)#public
)#class
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.