[R] pairwise.t.test non numeric factors error
peter dalgaard
pdalgd at gmail.com
Thu Jun 4 01:50:10 CEST 2015
> On 03 Jun 2015, at 22:05 , Erica Cseko Nolasco <ecnolasco at gmail.com> wrote:
>
> I tried using TRUE instead T, but it gave me the same error. (Script on the bottom)
> I really want a paired test and it seems to give me as we can see on the following matrix.
> No idea what is happening
>
I get the impression that you really do not know the difference between the paired t test and the independent samples t test. Please read up on it, it is off-topic here.
-pd
> > pairwise.t.test(x = data$tss,data$pa,p.adjust.method = 'bonferroni',paired=TRUE)
> Error in complete.cases(x, y) :
> not all arguments have the same length
>
> > pairwise.t.test(x = data$tss,data$algo,p.adjust.method = 'bonferroni')
>
> Pairwise comparisons using t tests with pooled SD
>
> data: data$tss and data$algo
>
> ANN CTA FDA GAM GBM GLM MARS MAXENT RF
> CTA 0.08118 - - - - - - - -
> FDA 0.00102 1.1e-11 - - - - - - -
> GAM 1.00000 1.00000 1.2e-05 - - - - - -
> GBM 0.01081 5.9e-10 1.00000 0.00021 - - - - -
> GLM 6.8e-06 3.8e-15 1.00000 3.3e-08 1.00000 - - - -
> MARS 0.16840 8.8e-08 1.00000 0.00620 1.00000 0.81420 - - -
> MAXENT 0.07340 1.00000 8.6e-12 1.00000 4.8e-10 3.0e-15 7.4e-08 - -
> RF 1.4e-06 3.3e-16 1.00000 5.3e-09 1.00000 1.00000 0.35873 2.6e-16 -
> SRE < 2e-16 < 2e-16 < 2e-16 < 2e-16 < 2e-16 < 2e-16 < 2e-16 < 2e-16 < 2e-16
>
> P value adjustment method: bonferroni
>
> Erica Csekö Nolasco
> Mestranda em Modelagem em Ciências da Terra e do Ambiente
> http://lattes.cnpq.br/2117508819823917
> Universidade Estadual de Feira de Santana
> Avenida Transnordestina s/n, Novo Horizonte
> Feira de Santana - BA, Brasil CEP 44.036-900.
>
> Graduate Student in Modeling of Environmental and Earth Sciences
> http://lattes.cnpq.br/2117508819823917
> Universidade Estadual de Feira de Santana
> Transnordestina Ave, Novo Horizonte
> Feira de Santana - BA, Brazil 44.036-900.
>
>
> 2015-06-03 16:55 GMT-03:00 peter dalgaard <pdalgd at gmail.com>:
>
> > On 03 Jun 2015, at 19:14 , Erica Cseko Nolasco <ecnolasco at gmail.com> wrote:
> >
> > Thanks Jim,
> >
> > I removed the corresponding cases and tried again. What I discovery now is
> > if I run the function without paired = T (pairwise.t.test(x =
> > data$tss,data$pa) is works. However, if I add the paired = T
> > (pairwise.t.test(x = data$tss,data$pa,paired = T) is gives me the error ' Error
> > in complete.cases(x, y) : not all arguments have the same length'. In
> > fact it seems to do a paired test even without the paired = T argument.
> > I really don't know why.
>
> What makes you say that? It would be a pretty bad error if it did paired tests and data aren't paired/matched. Are you sure that you actually want paired tests?
>
> -pd
>
>
> >
> > Regards,
> >
> > *Erica Csekö Nolasco*
> > Mestranda em Modelagem em Ciências da Terra e do Ambiente
> > http://lattes.cnpq.br/2117508819823917
> > Universidade Estadual de Feira de Santana
> > Avenida Transnordestina s/n, Novo Horizonte
> > Feira de Santana - BA, Brasil CEP 44.036-900.
> >
> > Graduate Student in Modeling of Environmental and Earth Sciences
> > http://lattes.cnpq.br/2117508819823917
> > Universidade Estadual de Feira de Santana
> > Transnordestina Ave, Novo Horizonte
> > Feira de Santana - BA, Brazil 44.036-900.
> >
> >
> > 2015-06-02 22:34 GMT-03:00 Jim Lemon <drjimlemon at gmail.com>:
> >
> >> Hi Erica,
> >> The problem may be that you are specifying a grouping factor (mdl) in
> >> which the group sizes are unequal. If one case in group "tot" is
> >> missing, is it possible to identify the corresponding cases in the
> >> other factor levels and delete them?
> >>
> >> Jim
> >>
> >>
> >> On Tue, Jun 2, 2015 at 11:59 PM, Erica Cseko Nolasco
> >> <ecnolasco at gmail.com> wrote:
> >>> Dear listers,
> >>>
> >>> I'm performing a PERMANOVA (adonis{vegan}) to compare the results (ROC,
> >>> TSS) of models based on two factors (model, algo). I was not able to
> >> find a
> >>> pairwise test for adonis, on PRIMER it would be a Tukey test. Though, I
> >>> chose to perform a pairwise.t.test what would be quite simple. However,
> >> no
> >>> matter I rearrange my response and factor vectors (as a factor or
> >> numeric)
> >>> it gives me the following error (the code on the bottom - Error in
> >>> complete.cases(x, y) :
> >>> not all arguments have the same length). I also tried to make a list of
> >>> the vectors, but it also gives me the error 'Error in sort.list(y) : 'x'
> >>> must be atomic for 'sort.list' Have you called 'sort' on a list?'
> >>>
> >>> I would appreciate any suggestions to solve this issue.
> >>>
> >>> Best,
> >>>
> >>> Erica
> >>>
> >>>> setwd('D:\\Erica\\mestrado\\analises\\results')
> >>>> library(vegan)
> >>>> rest=read.table('results_permanova.txt',sep="\t",header=T)
> >>>> rest$modelN=as.numeric(rest$model)
> >>>> rest$algoN=as.numeric(rest$algo)
> >>>> data=rest[complete.cases(rest),]
> >>>> head(data)
> >>> algo pa run model ROC TSS modelN algoN
> >>> 1 ANN PA1 RUN1 alt 0.947 0.867 2 2
> >>> 2 ANN PA10 RUN1 alt 0.978 0.869 2 2
> >>> 3 ANN PA11 RUN1 alt 0.993 0.931 2 2
> >>> 4 ANN PA12 RUN1 alt 0.961 0.845 2 2
> >>> 5 ANN PA13 RUN1 alt 0.988 0.960 2 2
> >>> 6 ANN PA14 RUN1 alt 0.996 0.988 2 2
> >>>> summary(data)
> >>> algo pa run model ROC
> >>> TSS
> >>> CTA :240 PA10 : 120 : 0 alt : 200 Min. :0.5290
> >>> Min. :0.0580
> >>> FDA :240 PA11 : 120 RUN1:2399 altPet : 200 1st Qu.:0.8780
> >>> 1st Qu.:0.7160
> >>> GAM :240 PA12 : 120 b1 : 200 Median :0.9350
> >>> Median :0.8270
> >>> GBM :240 PA13 : 120 b13 : 200 Mean :0.9184
> >>> Mean :0.7988
> >>> GLM :240 PA14 : 120 b13PETalt: 200 3rd Qu.:0.9790
> >>> 3rd Qu.:0.9110
> >>> MARS :240 PA15 : 120 b14 : 200 Max. :1.0000
> >>> Max. :1.0000
> >>> (Other):959 (Other):1679 (Other) :1199
> >>>
> >>> modelN algoN
> >>> Min. : 2.000 Min. : 2.000
> >>> 1st Qu.: 4.500 1st Qu.: 4.000
> >>> Median : 7.000 Median : 7.000
> >>> Mean : 7.498 Mean : 6.502
> >>> 3rd Qu.:10.000 3rd Qu.: 9.000
> >>> Max. :13.000 Max. :11.000
> >>>
> >>>> tss=data$TSS
> >>>> summary(tss)
> >>> Min. 1st Qu. Median Mean 3rd Qu. Max.
> >>> 0.0580 0.7160 0.8270 0.7988 0.9110 1.0000
> >>>> mdl=factor(data$model)
> >>>> summary(mdl)
> >>> alt altPet b1 b13 b13PETalt b14 b18
> >>> b7 pet pluv
> >>> 200 200 200 200 200 200 200
> >>> 200 200 200
> >>> temp tot
> >>> 200 199
> >>>> length(complete.cases(mdl,tss))
> >>> [1] 2399
> >>>> pairwise.t.test(tss,mdl,p.adj='bonf',paired=T,pool.sd = FALSE)
> >>> Error in complete.cases(x, y) :
> >>> not all arguments have the same length
> >>>>
> >>>
> >>> *Erica Csekö Nolasco*
> >>> Mestranda em Modelagem em Ciências da Terra e do Ambiente
> >>> http://lattes.cnpq.br/2117508819823917
> >>> Universidade Estadual de Feira de Santana
> >>> Avenida Transnordestina s/n, Novo Horizonte
> >>> Feira de Santana - BA, Brasil CEP 44.036-900.
> >>>
> >>> Graduate Student in Modeling of Environmental and Earth Sciences
> >>> http://lattes.cnpq.br/2117508819823917
> >>> Universidade Estadual de Feira de Santana
> >>> Transnordestina Ave, Novo Horizonte
> >>> Feira de Santana - BA, Brazil 44.036-900.
> >>>
> >>> [[alternative HTML version deleted]]
> >>>
> >>> ______________________________________________
> >>> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> >>> https://stat.ethz.ch/mailman/listinfo/r-help
> >>> PLEASE do read the posting guide
> >> http://www.R-project.org/posting-guide.html
> >>> and provide commented, minimal, self-contained, reproducible code.
> >>
> >
> > [[alternative HTML version deleted]]
> >
> > ______________________________________________
> > R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
> > https://stat.ethz.ch/mailman/listinfo/r-help
> > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> > and provide commented, minimal, self-contained, reproducible code.
>
> --
> Peter Dalgaard, Professor,
> Center for Statistics, Copenhagen Business School
> Solbjerg Plads 3, 2000 Frederiksberg, Denmark
> Phone: (+45)38153501
> Email: pd.mes at cbs.dk Priv: PDalgd at gmail.com
>
>
>
>
>
>
>
>
>
--
Peter Dalgaard, Professor,
Center for Statistics, Copenhagen Business School
Solbjerg Plads 3, 2000 Frederiksberg, Denmark
Phone: (+45)38153501
Email: pd.mes at cbs.dk Priv: PDalgd at gmail.com
More information about the R-help
mailing list