Package: D2MCS 1.0.1

Miguel Ferreiro-Díaz

D2MCS: Data Driving Multiple Classifier System

Provides a novel framework to able to automatically develop and deploy an accurate Multiple Classifier System based on the feature-clustering distribution achieved from an input dataset. 'D2MCS' was developed focused on four main aspects: (i) the ability to determine an effective method to evaluate the independence of features, (ii) the identification of the optimal number of feature clusters, (iii) the training and tuning of ML models and (iv) the execution of voting schemes to combine the outputs of each classifier comprising the Multiple Classifier System.

Authors:David Ruano-Ordás [aut, ctb], Miguel Ferreiro-Díaz [aut, cre], José Ramón Méndez [aut, ctb], University of Vigo [cph]

D2MCS_1.0.1.tar.gz
D2MCS_1.0.1.zip(r-4.5)D2MCS_1.0.1.zip(r-4.4)D2MCS_1.0.1.zip(r-4.3)
D2MCS_1.0.1.tgz(r-4.4-any)D2MCS_1.0.1.tgz(r-4.3-any)
D2MCS_1.0.1.tar.gz(r-4.5-noble)D2MCS_1.0.1.tar.gz(r-4.4-noble)
D2MCS_1.0.1.tgz(r-4.4-emscripten)D2MCS_1.0.1.tgz(r-4.3-emscripten)
D2MCS.pdf |D2MCS.html✨
D2MCS/json (API)
NEWS

# Install 'D2MCS' in R:
install.packages('D2MCS', repos = c('https://drordas.r-universe.dev', 'https://cloud.r-project.org'))

Peer review:

Bug tracker:https://github.com/drordas/d2mcs/issues

Uses libs:
  • openjdk– OpenJDK Java runtime, using Hotspot JIT

On CRAN:

3.70 score 191 downloads 70 exports 180 dependencies

Last updated 2 years agofrom:257090d1ba. Checks:OK: 5 NOTE: 2. Indexed: yes.

TargetResultDate
Doc / VignettesOKOct 11 2024
R-4.5-winNOTEOct 11 2024
R-4.5-linuxNOTEOct 11 2024
R-4.4-winOKOct 11 2024
R-4.4-macOKOct 11 2024
R-4.3-winOKOct 11 2024
R-4.3-macOKOct 11 2024

Exports:AccuracyBinaryPlotChiSquareHeuristicClassificationOutputClassMajorityVotingClassWeightedVotingClusterPredictionsCombinedMetricsCombinedVotingConfMatrixD2MCSDatasetDatasetLoaderDefaultModelFitDependencyBasedStrategyDependencyBasedStrategyConfigurationDIteratorExecutedModelsFisherTestHeuristicFIteratorFNFPGainRatioHeuristicGenericClusteringStrategyGenericHeuristicGenericModelFitGenericPlotHDDatasetHDSubsetInformationGainHeuristicKappaKendallHeuristicMCCMCCHeuristicMeasureFunctionMethodologyMinimizeFNMinimizeFPModelMultinformationHeuristicNoProbabilityNPVOddsRatioHeuristicPearsonHeuristicPPVPrecisionPredictionPredictionOutputProbAverageVotingProbAverageWeightedVotingProbBasedMethodologyRecallSensitivitySimpleStrategySimpleVotingSingleVotingSpearmanHeuristicSpecificityStrategyConfigurationSubsetSummaryFunctionTNTPTrainFunctionTrainOutputTrainsetTwoClassTypeBasedStrategyUseProbabilityVotingStrategy

Dependencies:askpassbase64encbitbit64brewbriobslibcachemcallrcaretclassclassIntclicliprclockcodetoolscolorspacecommonmarkcpp11crayoncredentialscurldata.tabledescdevtoolsdiagramdiffobjdigestdownlitdplyre1071ellipsisentropyevaluatefansifarverfastmapfontawesomeforcatsforeachfsFSelectorfuturefuture.applygenericsgertggplot2ggrepelghgitcredsglobalsgluegowergridExtragtablehardhathavenhighrhmshtmltoolshtmlwidgetshttpuvhttr2infotheoiniipredisobanditeratorsjquerylibjsonliteKernSmoothknitrlabelinglabelledlaterlatticelavalifecyclelistenvlubridatemagrittrMASSMatrixmccrmemoisemgcvmimeminiUImltoolsModelMetricsmunsellnlmennetnumDerivopensslparallellypillarpkgbuildpkgconfigpkgdownpkgloadplyrpraiseprettyunitspROCprocessxprodlimprofvisprogressprogressrpromisesproxypspurrrquestionrR.cacheR.methodsS3R.ooR.utilsR6raggrandomForestrappdirsrcmdcheckRColorBrewerRcppreadrrecipesrematch2remotesreshape2rJavarlangrmarkdownroxygen2rpartrprojrootrstudioapirversionsRWekaRWekajarssassscalessessioninfoshapeshinysourcetoolsSQUAREMstringistringrstylersurvivalsyssystemfontstestthattextshapingtibbletictoctidyrtidyselecttimechangetimeDatetinytextzdburlcheckerusethisutf8varhandlevctrsviridisLitevroomwaldowhiskerwithrxfunxml2xopenxtableyamlzip

A Brief Introduction to D2MCS

Miguel Ferreiro Diaz

Rendered fromD2MCS.Rmdusingknitr::rmarkdownon Oct 11 2024.

Last update: 2021-05-04
Started: 2021-04-28

Readme and manuals

Help Manual

Help pageTopics
Computes the Accuracy measure.Accuracy
Plotting feature clusters following bi-class problem.BinaryPlot
Feature-clustering based on ChiSquare method.ChiSquareHeuristic
D2MCS Classification Output.ClassificationOutput
Implementation of Majority Voting voting.ClassMajorityVoting
Implementation Weighted Voting scheme.ClassWeightedVoting
Manages the predictions achieved on a cluster.ClusterPredictions
Abstract class to compute the class prediction based on combination between metrics.CombinedMetrics
Implementation of Combined Voting.CombinedVoting
Confusion matrix wrapper.ConfMatrix
Data Driven Multiple Classifier System.D2MCS
Simple Dataset handler.Dataset
Dataset creation.DatasetLoader
Default model fitting implementation.DefaultModelFit
Clustering strategy based on dependency between features.DependencyBasedStrategy
Custom Strategy Configuration handler for the DependencyBasedStrategy strategy.DependencyBasedStrategyConfiguration
Feature-clustering based on Fisher's Exact Test.FisherTestHeuristic
Computes the False Negative errors.FN
Computes the False Positive value.FP
Feature-clustering based on GainRatio methodology.GainRatioHeuristic
Abstract Feature Clustering Strategy class.GenericClusteringStrategy
Abstract Feature Clustering heuristic object.GenericHeuristic
Abstract class for defining model fitting method.GenericModelFit
Pseudo-abstract class for creating feature clustering plots.GenericPlot
High Dimensional Dataset handler.HDDataset
High Dimensional Subset handler.HDSubset
Feature-clustering based on InformationGain methodology.InformationGainHeuristic
Computes the Kappa Cohen value.Kappa
Feature-clustering based on Kendall Correlation Test.KendallHeuristic
Computes the Matthews correlation coefficient.MCC
Feature-clustering based on Matthews Correlation Coefficient score.MCCHeuristic
Archetype to define customized measures.MeasureFunction
Abstract class to compute the probability prediction based on combination between metrics.Methodology
Combined metric strategy to minimize FN errors.MinimizeFN
Combined metric strategy to minimize FP errors.MinimizeFP
Feature-clustering based on Mutual Information Computation theory.MultinformationHeuristic
Compute performance across resamples.NoProbability
Computes the Negative Predictive Value.NPV
Feature-clustering based on Odds Ratio measure.OddsRatioHeuristic
Feature-clustering based on Pearson Correlation Test.PearsonHeuristic
Computes the Positive Predictive Value.PPV
Computes the Precision Value.Precision
Encapsulates the achieved predictions.PredictionOutput
Implementation of Probabilistic Average voting.ProbAverageVoting
Implementation of Probabilistic Average Weighted voting.ProbAverageWeightedVoting
Methodology to obtain the combination of the probability of different metrics.ProbBasedMethodology
Computes the Recall Value.Recall
Computes the Sensitivity Value.Sensitivity
Simple feature clustering strategy.SimpleStrategy
Abtract class to define simple voting schemes.SimpleVoting
Manages the execution of Simple Votings.SingleVoting
Feature-clustering based on Spearman Correlation Test.SpearmanHeuristic
Computes the Specificity Value.Specificity
Default Strategy Configuration handler.StrategyConfiguration
Classification set.Subset
Abstract class to computing performance across resamples.SummaryFunction
Computes the True Negative value.TN
Computes the True Positive Value.TP
Control parameters for train stage.TrainFunction
Stores the results achieved during training.TrainOutput
Trainning set.Trainset
Control parameters for train stage (Bi-class problem).TwoClass
Feature clustering strategy.TypeBasedStrategy
Compute performance across resamples.UseProbability
Voting Strategy template.VotingStrategy