RMixtCompUtilities

Description

MixtComp (Mixture Composer, https://github.com/modal-inria/MixtComp) is a model-based clustering package for mixed data originating from the Modal team (Inria Lille).

It has been engineered around the idea of easy and quick integration of all new univariate models, under the conditional independence assumption. Five basic models (Gaussian, Multinomial, Poisson, Weibull, NegativeBinomial) are implemented, as well as two advanced models (Func_CS and Rank_ISR). MixtComp has the ability to natively manage missing data (completely or by interval). MixtComp is used as an R package, but its internals are coded in C++ using state of the art libraries for faster computation.

This package contains plots, getters and format functions to simplify the use of RMixtComp and RMixtCompIO packages. It is recommended to use RMixtComp (instead of RMixtCompIO) which is more user-friendly.

Details

createAlgo gives you default values for required parameters.

convertFunctionalToVector, createFunctional and refactorCategorical functions help to transform data to the required format.

Getters are available to easily access some results: getBIC, getICL, getCompletedData, getParam, getTik, getEmpiricTik, getPartition, getType, getModel, getVarNames.

You can compute discriminative powers and similarities with functions: computeDiscrimPowerClass, computeDiscrimPowerVar, computeSimilarityClass, computeSimilarityVar.

Graphics functions are plot.MixtComp, heatmapClass, heatmapTikSorted, heatmapVar, histMisclassif, plotConvergence,plotDataBoxplot, plotDataCI, plotDiscrimClass, plotDiscrimVar, plotProportion.

See Also

RMixtComp RMixtCompIO Rmixmod packages

Available models

Description

Get information about models implemented in MixtComp

Usage

availableModels()

Value

a data.frame containing models implemented in MixtComp

model

model name

data.type

data type

format

Special format required for individuals

missing.formats

accepted formats (separated by a ;) for missing values

hyperparameter

Required hyperparameters in the paramStr elements of model object

comments

comments about the model

reference

link to article

Author(s)

Quentin Grimonprez

See Also

mixtCompLearn

Examples

availableModels()

Add the missing element to algo parameter

Description

Add the missing element to algo parameter with default values

Usage

completeAlgo(algo)

Arguments

Value

algo parameter with all required elements (see createAlgo function)

Author(s)

Quentin Grimonprez

Discriminative power

Description

Compute the discriminative power of each variable or class

Usage

computeDiscrimPowerVar(outMixtComp, class = NULL)

computeDiscrimPowerClass(outMixtComp)

Arguments

Details

The discriminative power of variable j is defined by 1 - C(j)

C(j)= - \sum_{k=1}^K sum_{i=1}^n P(Z_i=k|x_{ij}) \log(P(Z_i=k|x_{ij})) / (n*\log(K))

A high value (close to one) means that the variable is highly discriminating. A low value (close to zero) means that the variable is poorly discriminating.

The discriminative power of variable j in class k is defined by 1 - C(j)

C(j)= - sum_{i=1}^n (P(Z_i!=k|x_{ij}) \log(P(Z_i!=k|x_{ij})) + P(Z_i=k|x_{ij}) \log(P(Z_i=k|x_{ij}))) / (n*\log(2))

The discriminative power of class k is defined by 1 - D(k)

D(k) = -\sum_{i=1}^n P(Z_i=k|x_i) \log(P(Z_i=k|x_i)) / (n*\exp(-1))

Value

the discriminative power

Author(s)

Matthieu Marbac

See Also

plotDiscrimClass plotDiscrimVar

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)


  discVar <- computeDiscrimPowerVar(resLearn)
  discVarInClass1 <- computeDiscrimPowerVar(resLearn, class = 1)
  discClass <- computeDiscrimPowerClass(resLearn)

  # graphic representation of discriminant variables
  plotDiscrimVar(resLearn)
  # graphic representation of discriminant classes
  plotDiscrimClass(resLearn)
}

Similarity

Description

Compute the similarity between variables (or classes)

Usage

computeSimilarityVar(outMixtComp)

computeSimilarityClass(outMixtComp)

Arguments

Details

The similarities between variables j and h is defined by Delta(j,h)

Delta(j,h)^2 = 1 - \sqrt{(1/n) * \sum_{i=1}^n \sum_{k=1}^K (P(Z_i=k|x_{ij}) - P(Z_i=k|x_{ih}))^2}

The similarities between classes k and g is defined by 1 - Sigma(k,g)

Sigma(k,g)^2 = (1/n) * \sum_{i=1}^n (P(Z_i=k|x_i) - P(Z_i=g|x_i))^2

Value

a similarity matrix

Author(s)

Quentin Grimonprez

See Also

heatmapVar heatmapClass

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  simVar <- computeSimilarityVar(resLearn)
  simClass <- computeSimilarityClass(resLearn)
}

Convert a MixtComp functional string into a list of 2 vectors

Description

Convert a MixtComp functional string into a list of 2 vectors

Usage

convertFunctionalToVector(x)

Arguments

Value

a list of 2 vectors: time and value

Author(s)

Quentin Grimonprez

Examples

convertFunctionalToVector("1:5,1.5:12,1.999:2.9")

Create algo object

Description

create an algo object required by mixtCompLearn and mixtCompPredict from RMixtComp.

Usage

createAlgo(
  nbBurnInIter = 50,
  nbIter = 50,
  nbGibbsBurnInIter = 50,
  nbGibbsIter = 50,
  nInitPerClass = 50,
  nSemTry = 20,
  confidenceLevel = 0.95,
  ratioStableCriterion = 0.99,
  nStableCriterion = 20
)

Arguments

Value

a list with the parameters values

Author(s)

Quentin Grimonprez

Examples

# default values
algo <- createAlgo()

# change some values
algo <- createAlgo(nbIter = 200)

Create a functional in MixtComp format

Description

Create a functional in MixtComp format

Usage

createFunctional(time, value)

Arguments

Value

The functional data formatted to the MixtComp standard

Author(s)

Quentin Grimonprez

Examples

mat <- matrix(c(1, 2, 3, 9, 1, 1.5, 15, 1000), ncol = 2)
createFunctional(mat[, 1], mat[, 2])

Format the data parameter required by rmc

Description

format data.frame or matrix in list of character

Usage

formatData(data)

Arguments

Value

data as a list of characters

Author(s)

Quentin Grimonprez

Format the model parameter

Description

Format the model list for rmc/rmcMultiRun functions: - add paramStr when missing - ensure the list format of each element

Usage

formatModel(model)

Arguments

Value

model as a list where each element is the model applied to a variable (list with elements type and paramStr)

Author(s)

Quentin Grimonprez

Get criterion value

Description

Get criterion value

Usage

getBIC(outMixtComp)

getICL(outMixtComp)

Arguments

Value

value of the criterion

Author(s)

Quentin Grimonprez

See Also

Other getter: getCompletedData(), getEmpiricTik(), getMixtureDensity(), getParam(), getPartition(), getType()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # get criterion
  bic <- getBIC(resLearn)
  icl <- getICL(resLearn)
}

Get the completed data from MixtComp object

Description

Get the completed data from MixtComp object (does not manage functional models)

Usage

getCompletedData(outMixtComp, var = NULL, with.z_class = FALSE)

Arguments

Value

a matrix with the data completed by MixtComp (z_class is in the first column and then variables are sorted in alphabetic order, it may differ from the original order of the data).

Author(s)

Quentin Grimonprez

See Also

Other getter: getBIC(), getEmpiricTik(), getMixtureDensity(), getParam(), getPartition(), getType()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  # add missing values
  dataLearn$var1[12] <- "?"
  dataLearn$var2[72] <- "?"

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <- RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # get completedData
  completedData <- getCompletedData(resLearn)
  completedData2 <- getCompletedData(resLearn, var = "var1")
}

Get the tik

Description

Get the a posteriori probability to belong to each class for each individual

Usage

getEmpiricTik(outMixtComp)

getTik(outMixtComp, log = TRUE)

Arguments

Details

getTik returns a posteriori probabilities computed with the returned parameters. getEmpiricTik returns an estimation based on the sampled z_i during the algorithm.

Value

a matrix containing the tik for each individual (in row) and each class (in column).

Author(s)

Quentin Grimonprez

See Also

heatmapTikSorted

Other getter: getBIC(), getCompletedData(), getMixtureDensity(), getParam(), getPartition(), getType()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # get tik
  tikEmp <- getEmpiricTik(resLearn)
  tik <- getTik(resLearn, log = FALSE)
}

Get the mixture density

Description

Get the mixture density for each individual

Usage

getMixtureDensity(outMixtComp)

Arguments

Details

d(x_i) = \sum_k\pi_k P(x_i; \theta_k)

Value

a vector containing the mixture density for each individual.

Author(s)

Quentin Grimonprez

See Also

Other getter: getBIC(), getCompletedData(), getEmpiricTik(), getParam(), getPartition(), getType()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  d <- getMixtureDensity(resLearn)
}

Get the estimated parameter

Description

Get the estimated parameter

Usage

getParam(outMixtComp, var)

getProportion(outMixtComp)

Arguments

Value

the parameter of the variable

Author(s)

Quentin Grimonprez

See Also

plotDataBoxplot plotDataCI

Other getter: getBIC(), getCompletedData(), getEmpiricTik(), getMixtureDensity(), getPartition(), getType()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <- RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # get estimated parameters for variable var1
  param <- getParam(resLearn, "var1")
  prop <- getProportion(resLearn)
}

Get the estimated class from MixtComp object

Description

Get the estimated class from MixtComp object

Usage

getPartition(outMixtComp, empiric = FALSE)

Arguments

Value

a vector containing the estimated class for each individual.

Author(s)

Quentin Grimonprez

See Also

Other getter: getBIC(), getCompletedData(), getEmpiricTik(), getMixtureDensity(), getParam(), getType()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <- RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # get class
  estimatedClass <- getPartition(resLearn)
}

Names and Types Getters

Description

getType returns the type output of a MixtComp object, getModel returns the model object, getVarNames returns the name for each variable

Usage

getType(outMixtComp, with.z_class = FALSE)

getModel(outMixtComp, with.z_class = FALSE)

getVarNames(outMixtComp, with.z_class = FALSE)

Arguments

Value

a vector containing the type of models, names associated with each individual.

Author(s)

Quentin Grimonprez

See Also

Other getter: getBIC(), getCompletedData(), getEmpiricTik(), getMixtureDensity(), getParam(), getPartition()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # get type
  type <- getType(resLearn)

  # get model object
  model <- getModel(resLearn)

  # get variable names
  varNames <- getVarNames(resLearn)
}

Heatmap of the similarities between classes about clustering

Description

Heatmap of the similarities between classes about clustering

Usage

heatmapClass(output, pkg = c("ggplot2", "plotly"), ...)

Arguments

Details

The similarities between classes k and g is defined by 1 - Sigma(k,g)

Sigma(k,g)^2 = (1/n) * \sum_{i=1}^n (P(Z_i=k|x_i) - P(Z_i=g|x_i))^2

Author(s)

Matthieu MARBAC

See Also

computeSimilarityClass

Other plot: heatmapTikSorted(), heatmapVar(), histMisclassif(), plot.MixtComp(), plotConvergence(), plotDataBoxplot(), plotDataCI(), plotDiscrimClass(), plotDiscrimVar(), plotParamConvergence(), plotProportion()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # plot
  heatmapClass(resLearn)
}

Heatmap of the tik = P(Z_i=k|x_i)

Description

Heatmap of the tik = P(Z_i=k|x_i)

Usage

heatmapTikSorted(output, pkg = c("ggplot2", "plotly"), ...)

Arguments

Details

Observation are sorted according to the hard partition then for each component they are sorted by decreasing order of their tik's

Author(s)

Matthieu MARBAC

See Also

getTik

Other plot: heatmapClass(), heatmapVar(), histMisclassif(), plot.MixtComp(), plotConvergence(), plotDataBoxplot(), plotDataCI(), plotDiscrimClass(), plotDiscrimVar(), plotParamConvergence(), plotProportion()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # plot
  heatmapTikSorted(resLearn)
}

Heatmap of the similarities between variables about clustering

Description

Heatmap of the similarities between variables about clustering

Usage

heatmapVar(output, pkg = c("ggplot2", "plotly"), ...)

Arguments

Details

The similarities between variables j and h is defined by Delta(j,h)

Delta(j,h) = 1 - \sqrt{(1/n) * \sum_{i=1}^n \sum_{k=1}^K (P(Z_i=k|x_{ij}) - P(Z_i=k|x_{ih}))^2}

Author(s)

Matthieu MARBAC

See Also

computeSimilarityVar

Other plot: heatmapClass(), heatmapTikSorted(), histMisclassif(), plot.MixtComp(), plotConvergence(), plotDataBoxplot(), plotDataCI(), plotDiscrimClass(), plotDiscrimVar(), plotParamConvergence(), plotProportion()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # plot
  heatmapVar(resLearn)
}

Histogram of the misclassification probabilities

Description

Histogram of the misclassification probabilities

Usage

histMisclassif(output, pkg = c("ggplot2", "plotly"), ...)

Arguments

Details

Missclassification probability of observation i is denoted err_i err_i = 1 - max_k=1,...,K P(Z_i=k|x_i) Histograms of err_i's can be plotted for a specific class, all classes or every class

Author(s)

Matthieu MARBAC

See Also

Other plot: heatmapClass(), heatmapTikSorted(), heatmapVar(), plot.MixtComp(), plotConvergence(), plotDataBoxplot(), plotDataCI(), plotDiscrimClass(), plotDiscrimVar(), plotParamConvergence(), plotProportion()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # plot
  histMisclassif(resLearn)
}

Plot of a MixtComp object

Description

Plot of a MixtComp object

Usage

## S3 method for class 'MixtComp'
plot(
  x,
  nVarMaxToPlot = 3,
  pkg = c("ggplot2", "plotly"),
  plotData = c("CI", "Boxplot"),
  ...
)

Arguments

Author(s)

Quentin Grimonprez

See Also

mixtCompLearn mixtCompPredict

Other plot: heatmapClass(), heatmapTikSorted(), heatmapVar(), histMisclassif(), plotConvergence(), plotDataBoxplot(), plotDataCI(), plotDiscrimClass(), plotDiscrimVar(), plotParamConvergence(), plotProportion()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  plot(resLearn)
}

Convergence of algorithm

Description

Plot the evolution of the completed loglikelihood during the SEM algorithm. The vertical line denotes the end of the burn-in phase.

Usage

plotConvergence(output, ...)

Arguments

Details

This function can be used to check the convergence and choose the parameters nbBurnInIter and nbIter from mcStrategy.

Author(s)

Quentin Grimonprez

See Also

Other plot: heatmapClass(), heatmapTikSorted(), heatmapVar(), histMisclassif(), plot.MixtComp(), plotDataBoxplot(), plotDataCI(), plotDiscrimClass(), plotDiscrimVar(), plotParamConvergence(), plotProportion()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # plot
  plotConvergence(resLearn)
}

Boxplot per class

Description

Display a boxplot (5

Usage

plotDataBoxplot(
  output,
  var,
  class = seq_len(output$algo$nClass),
  grl = TRUE,
  pkg = c("ggplot2", "plotly"),
  ...
)

Arguments

Details

For functional data, three other parameters are available:

add.obs

if TRUE, observations are added to the plot. Default = FALSE.

ylim

ylim of the plot.

xlim

xlim of the plot.

Author(s)

Matthieu MARBAC

See Also

Other plot: heatmapClass(), heatmapTikSorted(), heatmapVar(), histMisclassif(), plot.MixtComp(), plotConvergence(), plotDataCI(), plotDiscrimClass(), plotDiscrimVar(), plotParamConvergence(), plotProportion()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # plot
  plotDataBoxplot(resLearn, "var1")
}

Mean and 95%-level confidence intervals per class

Description

Mean and 95%-level confidence intervals per class

Usage

plotDataCI(
  output,
  var,
  class = seq_len(output$algo$nClass),
  grl = FALSE,
  pkg = c("ggplot2", "plotly"),
  ...
)

Arguments

Details

For functional data, three other parameters are available:

add.obs

if TRUE, observations are added to the plot. Default = FALSE.

add.CI

if FALSE, confidence intervals are removed from the plot. Default = TRUE.

xlim

xlim of the plot.

ylim

ylim of the plot.

Author(s)

Matthieu MARBAC

See Also

Other plot: heatmapClass(), heatmapTikSorted(), heatmapVar(), histMisclassif(), plot.MixtComp(), plotConvergence(), plotDataBoxplot(), plotDiscrimClass(), plotDiscrimVar(), plotParamConvergence(), plotProportion()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # plot
  plotDataCI(resLearn, "var1")
}

Barplot of the discriminative power of the classes

Description

Barplot of the discriminative power of the classes

Usage

plotDiscrimClass(output, ylim = c(0, 1), pkg = c("ggplot2", "plotly"), ...)

Arguments

Details

The discriminative power of class k is defined by 1 - D(k)

D(k) = -\sum_{i=1}^n P(Z_i=k|x_i) \log(P(Z_i=k|x_i)) / (n*\exp(-1))

Author(s)

Matthieu MARBAC

See Also

computeDiscrimPowerClass

Other plot: heatmapClass(), heatmapTikSorted(), heatmapVar(), histMisclassif(), plot.MixtComp(), plotConvergence(), plotDataBoxplot(), plotDataCI(), plotDiscrimVar(), plotParamConvergence(), plotProportion()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  plotDiscrimClass(resLearn)
}

Barplot of the discriminative power of the variables

Description

Barplot of the discriminative power of the variables

Usage

plotDiscrimVar(
  output,
  class = NULL,
  ylim = c(0, 1),
  pkg = c("ggplot2", "plotly"),
  ...
)

Arguments

Details

The discriminative power of variable j is defined by 1 - C(j)

C(j)= -\sum_{k=1}^K \sum_{i=1}^n P(Z_i=k|x_{ij}) ln(P(Z_i=k|x_{ij})) / (n*\log(K))

Author(s)

Matthieu MARBAC

See Also

computeDiscrimPowerVar

Other plot: heatmapClass(), heatmapTikSorted(), heatmapVar(), histMisclassif(), plot.MixtComp(), plotConvergence(), plotDataBoxplot(), plotDataCI(), plotDiscrimClass(), plotParamConvergence(), plotProportion()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # plot
  plotDiscrimVar(resLearn)

  plotDiscrimVar(resLearn, class = 1)
}

Evolution of parameters

Description

Plot the evolution of estimated parameters after the burn-in phase.

Usage

plotParamConvergence(output, var, ...)

Arguments

Author(s)

Quentin Grimonprez

See Also

Other plot: heatmapClass(), heatmapTikSorted(), heatmapVar(), histMisclassif(), plot.MixtComp(), plotConvergence(), plotDataBoxplot(), plotDataCI(), plotDiscrimClass(), plotDiscrimVar(), plotProportion()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # plot
  plotParamConvergence(resLearn, "var1")
  plotParamConvergence(resLearn, "var2")
}

Plot the mixture's proportions

Description

Plot the mixture's proportions

Usage

plotProportion(output, pkg = c("ggplot2", "plotly"), ...)

Arguments

Author(s)

Quentin Grimonprez

See Also

Other plot: heatmapClass(), heatmapTikSorted(), heatmapVar(), histMisclassif(), plot.MixtComp(), plotConvergence(), plotDataBoxplot(), plotDataCI(), plotDiscrimClass(), plotDiscrimVar(), plotParamConvergence()

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  # plot
  plotProportion(resLearn)
}

Print Values

Description

Print a MixtComp object

Usage

## S3 method for class 'MixtComp'
print(x, nVarMaxToPrint = 5, ...)

Arguments

Author(s)

Quentin Grimonprez

See Also

mixtCompLearn mixtCompPredict

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  print(resLearn)
}

Rename a categorical value

Description

Rename a categorical value

Usage

refactorCategorical(
  data,
  oldCateg = unique(data),
  newCateg = seq_along(oldCateg)
)

Arguments

Value

Data with new categorical values

Author(s)

Quentin Grimonprez

Examples

dat <- c("single", "married", "married", "divorced", "single")
refactorCategorical(dat, c("single", "married", "divorced"), 1:3)

MixtComp Object Summaries

Description

Summary of a MixtComp object

Usage

## S3 method for class 'MixtComp'
summary(object, ...)

Arguments

Author(s)

Quentin Grimonprez

See Also

mixtCompLearn print.MixtComp

Examples

if (requireNamespace("RMixtCompIO", quietly = TRUE)) {
  dataLearn <- list(
    var1 = as.character(c(rnorm(50, -2, 0.8), rnorm(50, 2, 0.8))),
    var2 = as.character(c(rnorm(50, 2), rpois(50, 8)))
  )

  model <- list(
    var1 = list(type = "Gaussian", paramStr = ""),
    var2 = list(type = "Poisson", paramStr = "")
  )

  algo <- list(
    nClass = 2,
    nInd = 100,
    nbBurnInIter = 100,
    nbIter = 100,
    nbGibbsBurnInIter = 100,
    nbGibbsIter = 100,
    nInitPerClass = 3,
    nSemTry = 20,
    confidenceLevel = 0.95,
    ratioStableCriterion = 0.95,
    nStableCriterion = 10,
    mode = "learn"
  )

  resLearn <-RMixtCompIO::rmcMultiRun(algo, dataLearn, model, nRun = 3)

  summary(resLearn)
}