R语言生存分析大全Word文档下载推荐.docx

R语言生存分析大全Word文档下载推荐.docx

《R语言生存分析大全Word文档下载推荐.docx》由会员分享，可在线阅读，更多相关《R语言生存分析大全Word文档下载推荐.docx（18页珍藏版）》请在冰豆网上搜索。

packagecomputestheKaplan-Meierestimatorfortruncatedand/orcensoreddata. 

rms 

（replacementoftheDesignpackage）proposesamodifiedversionofthe 

function.The 

prodlim 

packageimplementsafastalgorithmandsomefeaturesnotincludedin 

survival.VariousconfidenceintervalsandconfidencebandsfortheKaplan-Meierestimatorareimplementedinthe 

km.ci 

package. 

plot.Surv 

ofpackage 

eha 

plotstheKaplan-Meierestimator.The 

NADA 

packageincludesafunctiontocomputetheKaplan-Meierestimatorforleft-censoreddata. 

svykm 

in 

survey 

providesaweightedKaplan-Meierestimator. 

nested.km 

NestedCohort 

estimatesthesurvivalcurveforeachlevelofcategoricalvariableswithmissingdata.The 

kaplan-meier 

functionin 

spatstat 

computestheKaplan-Meierestimatorfromhistogramdata.The 

MAMSE 

packagepermitstocomputeaweightedKaplan-Meierestimate.The 

KM 

functioninpackage 

rhosp 

plotsthesurvivalfunctionusingavariantoftheKaplan-Meierestimatorinahospitalisationriskcontext.The 

survPresmooth 

packagecomputespresmoothedestimatesofthemainquantitiesusedforright-censoreddata,i.e.,survival,hazardanddensityfunctions.The 

asbio 

packagepermitstocomputetheKaplan-MeierestimatorfollowingPollocketal.（1998）.The 

bpcp 

packageprovidesseveralfunctionsforcomputingconfidenceintervalsofthesurvivaldistribution（e.g.,betaproductconfidenceprocedure）.The 

lbiassurv 

packageoffersvariouslength-biascorrectionstosurvivalcurveestimation.Non-ParametricconfidancebandsfortheKaplan-Meierestimatorcanbecomputedusingthe 

kmconfband 

package.The 

kmc 

packageimplementstheKaplan-Meierestimatorwithconstraints.

∙Non-Parametricmaximumlikelihoodestimation（NPMLE）:

Icens 

packageprovidesseveralwaystocomputetheNPMLEofthesurvivaldistributionforvariouscensoringandtruncationschemes.MLEcens 

canalsobeusedtocomputetheMLEforinterval-censoreddata. 

dblcens 

permitstocomputetheNPMLEofthecumulativedistributionfunctionforleft-andright-censoreddata.The 

icfitfunctioninpackage 

interval 

computestheNPMLEforinterval-censoreddata.The 

DTDA 

implementsseveralalgorithmspermittingtoanalysepossiblydoublytruncatedsurvivaldata.

∙Parametric:

fitdistrplus 

packagepermitstofitanunivariatedistributionbymaximumlikelihood.Datacanbeintervalcensored.The 

vitality 

packageprovidesroutinesforfittingmodelsinthevitalityfamilyofmortalitymodels.

HazardEstimation

∙The 

muhaz 

packagepermitstoestimatethehazardfunctionthroughkernelmethodsforright-censoreddata.

epi.insthaz 

functionfrom 

epiR 

computestheinstantaneoushazardfromtheKaplan-Meierestimator.

∙polspline, 

gss 

and 

logspline 

allowtoestimatethehazardfunctionusingsplines.

ICE 

packageaimsatestimatingthehazardfunctionforintervalcensoreddata.

bshazard 

packageprovidesnon-parametricsmoothingofthehazardthroughB-splines.

Testing

survdiff 

comparessurvivalcurvesusingtheFleming-HarringtonG-rhofamilyoftest. 

implementsthisclassoftestsforleft-censoreddata.

∙clinfun 

implementsapermutationversionofthelogranktestandaversionofthelogrankthatadjustsforcovariates.

exactRankTests 

implementstheshift-algorithmbyStreitbergandRoehmelforcomputingexactconditionalp-valuesandquantiles,possiblyforcensoreddata.

∙SurvTest 

inthe 

coin 

packageimplementsthelogranktestreformulatedasalinearranktest.

maxstat 

packageperformstestsusingmaximallyselectedrankstatistics.

packageimplementslogrankandWilcoxontypetestsforinterval-censoreddata.

∙Threegeneralisedlogranktestsandascoretestforinterval-censoreddataareimplementedinthe 

glrt 

package.

∙survcomp 

compares2hazardratios.

TSHRC 

implementsatwostageprocedureforcomparinghazardfunctions.

Survgini 

packageproposestotesttheequalityoftwosurvivaldistributionsbasedontheGiniindex.

FHtest 

packageoffersseveraltestsbasedontheFleming-Harringtonclassforcomparingsurivalcurveswithright-andinterval-censoreddata.

LogrankA 

packageprovidesalogranktestforwhichaggregateddatacanbeusedasinput.

∙Theshorttermandlongtermhazardratiomodelfortwosamplessurvivaldatacanbefoundinthe 

YPmodel 

RegressionModelling

∙Coxmodel:

coxph 

functioninthe 

packagefitstheCoxmodel. 

cph 

packageandthe 

packageproposesomeextensionstothe 

function.Thepackage 

coxphf 

implementstheFirth'

spenalisedmaximumlikelihoodbiasreductionmethodfortheCoxmodel.AnimplementationofweightedestimationinCoxregressioncanbefoundin 

coxphw.The 

coxrobust 

packageproposesarobustimplementationoftheCoxmodel. 

timecox 

inpackage 

timereg 

fitsCoxmodelswithpossiblytime-varyingeffects.The 

mfp 

packagepermitstofitCoxmodelswithmultiplefractionalpolynomial.The 

fitsCoxmodelsforcovariateswithmissingdata.ACoxmodelmodelcanbefittedtodatafromcomplexsurveydesignusingthe 

svycoxph 

survey.The 

multipleNCC 

packagefitsCoxmodelsusingaweightedpartiallikelihoodfornestedcase-controlstudies.The 

intcox 

packageimplementstheCoxmodelforinterval-censoreddatausingtheICM-algorithm.The 

MIICD 

packageimplementsPan'

s（2000）multipleimputationapproachtoCoxmodelsforintervalcensoreddata.The 

ICsurv 

packagefitsCoxmodelsforinterval-censoreddatathroughanEMalgorithm.The 

dynsurvpackagefitstime-varyingcoefficientmodelsforintervalcensoredandrightcensoredsurvivaldatausingaBayesianCoxmodel,asplinebasedCoxmodeloratransformationmodel.The 

CPHshapepackagecomputestheCoxproportionalhazardsmodelwithshapeconstrainedhazardfunctions.The 

OrdFacReg 

packageimplementstheCoxmodelusinganactivesetalgorithmfordummyvariablesoforderedfactors.The 

survivalMPL 

packagefitsCoxmodelsusingmaximumpenalisedlikelihoodandprovideanonparametricsmoothestimateofthebaselinehazardfunction. 

cumres 

gof 

computesgoodness-of-fitmethodsfortheCoxproportionalhazardsmodel.Theproportionalityassumptioncanbecheckedusingthe 

cox.zph 

survival.The 

CPEpackagecalculatesconcordanceprobabilityestimatefortheCoxmodel,asdoesthe 

coxphCPE 

clinfun.The 

coxphQuantile 

inthelatterpackagedrawsaquantilecurveofthesurvivaldistributionasafunctionofcovariates.The 

multcomp 

packagecomputessimultaneoustestsandconfidenceintervalsfortheCoxmodelandotherparametricsurvivalmodels.The 

lsmeans 

packagepermitstoobtainleast-squaresmeans（andcontraststhereof）fromlinearmodels.Inparticular,itprovidessupportforthe 

coxph, 

survreg 

coxme 

functions.The 

multtest 

packageonBioconductorproposesaresamplingbasedmultiplehypothesistestingthatcanbeappliedtotheCoxmodel.TestingcoefficientsofCoxregressionmodelsusingaWaldtestwithasandwichestimatorofvariancecanbedoneusingthe 

sawspackage.The 

rankhazard 

packagepermitstoplotvisualisationoftherelativeimportanceofcovariatesinaproportionalhazardsmodel.The 

smoothHR 

packageprovideshazardratiocurvesthatallowsfornonlinearrelationshipbetweenpredictorandsurvival.The 

paf 

packagepermitstocomputetheunadjusted/adjustedattributablefractionfunctionfromaCoxproportionalhazardsmodel.The 

PHevalpackageproposestoolstochecktheproportionalhazardsassumptionusingastandardisedscoreprocess.

∙ParametricProportionalHazardsModel:

（from 

survival）fitsaparametricproportionalhazardsmodel.The 

mixPHM 

packagesimplementaproportionalhazardsmodelwithaparametricbaselinehazard.The 

pphsm 

translatesanAFTmodeltoaproportionalhazardsform.The 

polspline 

packageincludesthe 

hare 

functionthatfitsahazardregressionmodel,usingsplinestomodelthebaselinehazard.Hazardscanbe,butnotnecessarily,proportional.The 

flexsurv 

packageimplementsthemodelofRoystonandParmar（2002）.Themodelusesnaturalcubicsplinesforthebaselinesurvivalfunction,andproportionalhazards,proportionaloddsorprobitfunctionsforregression.

∙AcceleratedFailureTime（AFT）Models:

canfitanacceleratedfailuretimemodel.Amodifiedversionof 

isimplementedinthe 

package（ 

psmfunction）.Itpermitstousesomeofthe 

functionalities.The 

packagealsoproposesanimplementationoftheAFTmodel（function 

aftreg）.AnAFTmodelwithanerrordistributionassumedtobeamixtureofG-splinesisimplementedinthe 

smoothSurv 

packageproposesthefrontendofthe 

functionforleft-censoreddata.Aleast-squareprincipledimplementationoftheAFTmodelcanbefoundinthe 

lss 

simexaft 

packageimplementstheSimulation-ExtrapolationalgorithmfortheAFTmodel,thatcanbeusedwhencovariatesaresubjecttomeasurementerror.Arobustversionoftheacceleratedfailuretimemodelcanbef