Movatterモバイル変換

diceR 3.1.0

• Remove deprecated helper functions packageclvcurrently pending archival
• Improve documentation for parameterk.method inconsensus_evaluation()

diceR 3.0.0

• There were errors running tests on clang-UBSAN and gcc-UBSAN builds,due to usage of functions fromclusterSim. Reinstateinternal and external functions fromclusterCrit package,which is no longer archived.
• Use new parameterverbose indice() tocontrol console printouts of main tasks being performed instead of usingprogress
• Refactorgraph_heatmap() to usepheatmap::pheatmap() sinceNMF::aheatmap()throws agridPLT() error whenever it is run in a script orinteractively, but not in R markdown documents. Thus there are run timeerrors when used in unit tests and examples. See renozao/NMF#65
• Remove deprecated function usage

diceR 2.2.0

• Newabs argument inconsensus_cluster():control whether to apply absolute value to Spearman and Pearsoncorrelation matrices before subtracting from one (@tiagochst, #161)
• New distance matrix option,distance = "pearson"
• Useblockcluster instead ofmixedClust asthe latter is now archived
• Update tests inconsensus_cluster() that skip whensuggested packages are not installed

diceR 2.1.0

• Sort cluster sizesk correctly in relative change inarea under CDF curve (@IgnatiusPang, #167) and consensusmatrix CDF graphs
• Replace deprecatedaes_ calls with tidyeval idioms
• Passp.item andseed.data arguments todice() (#162, #165)

diceR 2.0.0

Internal and external validity indices were refactored to avoid usinghelper functions from theclusterCrit package, which isscheduled to be archived. Please adapt your code if it extractsdeprecated validity indices, as described below.

• Calinski-Harabasz index now calculated usingclusterSim::index.G1()
• Dunn index now calculated usingclValid::dunn()
• Gamma index now calculated usingclusterSim::index.G2()
• C-index now calculated usingclusterSim::index.C()
• Davies-Bouldin now calculated usingclusterSim::index.DB()
• SD index now calculatedcorrectly usingclv::clv.SD() and helper functions fromclv.Previously only the total separation between clusters was returned.
• S_Dbw index now calculated usingclv::clv.SDbw() andhelper functions fromclv
• Rousseeuw’s Silhouette now calculated usingclusterSim::index.S()
• PBM, Tau, McClain-Rao, Ray-Turi, and G-plus indices were removed asequivalent implementations from other packages were not found. They maybe reinstated in the future.
• All external validity indices are now calculated manually usingcounts from the concordance matrix (Hubert, Jaccard, McNemar, and Randindices)
• Best index value (maximum or minimum) now calculated manually

diceR 1.2.2

• Passlower andupper parameters fromPAC() toconsensus_evaluate() (#160)

diceR 1.2.1

• References added fork_modes() andCSPA()consensus functions (#157)
• Update roxygen2 to avoid HTML5 documentation notes
• UsemixedClust instead ofblockcluster forco-clustering since the latter keeps getting archived

diceR 1.2.0

• Usetestthat::skip_if_not_installed() to run testsconditionally when using packages in Suggests
• Use roxygen tag@examplesIf andrlang::is_installed() to run examples conditionally whenusing packages in Suggests

diceR 1.1.0

• Reinstateblockcluster package as it is under activemaintenance again

diceR 1.0.4

• Suppress new names messages from transformed NMF data

• Flattened matrices include 4th dimension of clusteringarray

diceR 1.0.3

• Add package logo usinghexSticker

• Packageblockcluster has been archived, remove fromSuggests

diceR 1.0.2

• Remove deprecatedcontext() in tests

• AddRColorBrewer to Suggests because it is crossreferenced in?graphs

• Add trailing slash for URLs in DESCRIPTION

• Removetibble from Imports, no longer used

diceR 1.0.1

• Suppress warnings whenclValid::connectivity() iscalled regarding comparison with more than one class. Since R-4.0.0, amatrix object also inherits from classarray

• Inalgii_heatmap(), the objectiialready has row names passed from columns which are named vectors (issuealso identified in #148, thanks@romainfrancois)

diceR 1.0.0

Decreased dependencies

The following steps were taken to minimize dependencies and ensurethatdiceR can still run on R 3.5:

• Removedcli andRColorBrewer fromImports

• Movedapcluster,blockcluster,cluster,dbscan,e1071,kernlab, andkohonen toSuggests,use their specific clustering algorithms conditionally.mclust needs to be inImports becausemclust::mclustBIC() needs to be imported

• Movedsigclust toSuggests, use withinsigclust() conditionally

• Movedprogress toSuggests, use withinconsensus_cluster() conditionally

• MovedpoLCA toSuggests, use withindice() conditionally

• MovedRtsne toSuggests, use withinprepare_data() conditionally

• Removed old dependencygrDevices fromImports

• Set minimum version to R (>= 3.5) forklaRdependencyquestionr

• Inev_confmat(), useyardstick::conf_mat() instead ofcaret::confusionMatrix().caret has manydependencies, so best to avoid using it

• Ingraph_heatmap(), useNMF::aheatmap()instead ofgplots::heatmap.2().gplots dependsoncaTools, which now relies on R (>= 3.6)

• Inconsensus_cluster(), usestringr::str_to_title() instead ofHmisc::capitalize().Hmisc depends onlatticeExtra, which now relies on R (>= 3.6)

• Ingraph_delta_area(), use base solution instead offlux::auc().flux also depends oncaTools

• Inprepare_data(), use own implementation ofquantable::robustscale() with all of the former function’sdefaults.quantable also depends oncaTools

• Specify Bioconductor installation on Travis and AppVeyor sinceNMF now ImportsBiobase

Minor improvements and bugfixes

• RemovesuppressWarnings(RNGversion("3.5.0")) afterupdating R version

• RunLCA() unit test on imputed clusteringobject

• Remove internal validity measures with anyInfentries forconsensus_reweigh()

• Use a cleaner, more robust method of removingRplots.pdf after runningtest-graphs.R

• Ensure column binding withpurrr::map_dfc() inconsensus_rank()

• Replaceddplyr::bind_cols() withpurrr::flatten_dfc() to suppress warning “Outer names areonly allowed for unnamed scalar” inget_cdf()

• update roxygen and docs

diceR 0.6.0

• Remove deprecateddplyr functions and use.data pronoun

• k-means clustering should not support distance matrices as input(@jerryji1993,#139)

• Add LCA as a consensus function (@philstraforelli, #137)

diceR 0.5.2

• Fixlength > 1 in coercion to logical error inconsensus_evaluate() due to comparisons using|| operator

• AddsuppressWarnings(RNGversion("3.5.0")) beforecall toset.seed() in examples, tests, and vignette to useold RNG sampling

• Use.covrignore to excludezzz.R frombeing considered in code coverage

• Usedplyr version >= 0.7.5 to ensurebind_rows() works

• Fixed bug where scaled matrix using the “robust” method inprepare_data() was nested in thedata element(@AlineTalhouk,#134)

diceR 0.5.1

• Add parameterhc.method indice andconsensus_cluster to pass tomethod parameterinstats::hclust (@JakeNel28, #130)

• Remove dependencies onlargeVis: package will bearchived

diceR 0.5.0

• Revert back to usingNMF sinceNNLM hasbeen archived andNMF is back in activemaintenance.

• Choose fuzzifier m incmeans using Equation 5 fromhttps://academic.oup.com/bioinformatics/article/26/22/2841/227572(thanks@Asduveneck)

diceR 0.4.0

• Replace all code that depended onNMF withNNLM andpheatmap: CRAN notified thatNMF will be archived because of inactivemaintenance

• Update.yml files default templates

diceR 0.3.2

• Fix bug inconsensus_cluster() when customalgorithms were excluded from output (thanks@phiala)

• Use markdown language for documentation

• Various performance improvements and codesimplifications

diceR 0.3.1

• Suppress success/fail message printout and fix input data to bematrix for block clustering

• Fix bug inalgii_heatmap() whenk.method = "all" indice()

• Fix bug in calculating internal indices when data has categoricalvariables (thanks Kurt Salmela)

diceR 0.3.0

• Updated object output names inconsensus_evaluate()

• Fix unit test intest-dice.R for R-devel

• Add internal function: ranked algorithms vs internal validityindices heatmap graph

• Fix bugs ingraph_cdf(),graph_tracking() when only one k selected

• Progress messages indice()

• Fix bug inconsensus_evaluate() when algorithm hasNA for all PAC values

diceR 0.2.0

• New dimension reduction methods: t-SNE, largeVis (@dustin21)

• Better annotated progress bar usingprogresspackage

• Speed up the operation that transforms a matrix to become“NMF-ready”

• Simplify saving mechanism inconsensus_cluster()such that onlyfile.name needs to be specified, and thesave parameter has been removed

• New algorithms: SOM, Fuzzy C-Means, DBSCAN (@dustin21, #118)

• Added significance testing section to vignette

• Fixed direction of optimization: compactness and connectivityshould be minimized

diceR 0.1.0

• New submission to CRAN accepted on June 21, 2017