Re: [R] how to recover a list structure
Here is one way that might work. x <- list(a=runif(10), b=runif(30), c=runif(25)) # unlist and then construct a factor to put them back together x.u <- unlist(x) x.f <- unlist(mapply(rep, names(x), sapply(x, length))) # find the lowest 5 values and set to -1 x.u[order(x.u)[1:5]] <- -1 # put back together split(x.u, x.f) On Sat, Feb 21, 2009 at 11:51 PM, wrote: > I am experiencing some problems at working with lists at high level. > In the following "coef" contains the original DWT coefficients organized in a > list. > Thorugh applying the following two commands: > coef.abs <- lapply(unlist(coef,recursive=FALSE,use.names =TRUE),abs) > coef.abs.sorted <- sort(unlist(coef.abs),decreasing=TRUE) > I get vector "coef.abs.sorted" containing the coefficients absolute value > ordered by decreasing magnitude, whose first 10 elements I set to zero after > applying an iterative math formula (telling me how many coefficient values to > spare). > My problem is setting to zero the correspondent coefficients in the original > "coef" list so as to reconstruct the signal from the depleted coefficients > set, given that the above operations have erased the list structure. I can > get the labels of the coefficients that have to be zeroed ... but the names > of vector "coef.abs.sorted" elements differes from the original "coef" list > elements as shown in the following: > > names(coef.abs.sorted[1:10]) > [1] "d6.d6(0)" "d5.d5(2)" "d5.d5(0)" "d4.d4(2)" "d5.d5(1)" "d4.d4(1)" > "d4.d4(6)" "d4.d4(5)" "d4.d4(3)" > [10] "d3.d3(12)" > I wonder if there exists a high level command to set to zero respectively the > coefficients in list "coef": > - Level=6, Oscillation-Number=0 > - Level=5, Oscillation-Number=0,1,2 > - Level=4, Oscillation-Number=1,2,3,5.6 > - Level=3, Oscillation-Number=12 > > >> coef > $d1 >d1(0) d1(1) d1(2) d1(3) d1(4) > d1(5) d1(6) d1(7) > 1.075111e-02 -4.777441e-03 -2.775026e-03 2.011722e-03 -2.376611e-02 > -1.574422e-03 -2.810372e-02 8.542393e-03 >d1(8) d1(9)d1(10)d1(11)d1(12) > d1(13)d1(14)d1(15) > -1.754845e-02 -7.363856e-04 3.123922e-03 1.053743e-02 -1.014095e-02 > 1.059210e-02 -4.152311e-03 -7.320749e-04 > d1(16)d1(17)d1(18)d1(19)d1(20) > d1(21)d1(22)d1(23) > -3.680064e-03 5.620248e-03 -2.251741e-03 -1.454674e-02 1.891275e-02 > 2.092009e-03 -1.286729e-02 9.037282e-03 > d1(24)d1(25)d1(26)d1(27)d1(28) > d1(29)d1(30)d1(31) > -1.027137e-02 1.014230e-02 -2.175260e-03 -1.689383e-03 -2.016800e-03 > 3.984632e-03 -1.563781e-03 -3.801979e-03 > d1(32)d1(33)d1(34)d1(35)d1(36) > d1(37)d1(38)d1(39) > 6.976144e-04 1.048419e-03 -1.319865e-02 -8.368520e-04 6.618890e-03 > -5.835673e-03 8.721318e-03 2.018621e-02 > d1(40)d1(41)d1(42)d1(43)d1(44) > d1(45)d1(46)d1(47) > -3.329934e-03 -4.823901e-03 2.957486e-02 -2.887600e-02 2.500062e-02 > -2.207839e-02 1.516034e-02 4.328945e-05 > d1(48)d1(49)d1(50)d1(51)d1(52) > d1(53)d1(54)d1(55) > 1.605634e-02 -2.360298e-02 7.875788e-03 -4.129065e-03 -1.014460e-02 > -2.676679e-03 -9.953614e-03 -8.844977e-03 > d1(56)d1(57)d1(58)d1(59) > -2.141513e-02 -7.405836e-03 -5.078857e-03 8.480675e-04 > > $d2 > d2(0)d2(1)d2(2)d2(3)d2(4)d2(5) > d2(6)d2(7)d2(8) > 0.033393680 0.015767682 -0.006008297 0.016899336 0.006661543 0.005297962 > -0.005782728 -0.009358879 -0.012128449 > d2(9) d2(10) d2(11) d2(12) d2(13) d2(14) > d2(15) d2(16) d2(17) > -0.014097771 0.003439004 -0.010154074 0.031203264 -0.057290615 0.004552577 > 0.061358997 -0.040730889 0.004064522 > d2(18) d2(19) d2(20) d2(21) d2(22) d2(23) > d2(24) d2(25) d2(26) > -0.007208508 0.010426952 0.003855540 -0.011492580 0.035028591 0.029604957 > -0.020069126 -0.018983905 -0.033710992 > d2(27) d2(28) d2(29) > 0.035848240 -0.014716622 0.008372893 > > $d3 > d3(0)d3(1)d3(2)d3(3)d3(4)d3(5) > d3(6)d3(7)d3(8) > 0.020437693 -0.023658475 0.033448175 0.009689787 0.024750901 0.040022643 > 0.016532575 -0.035730498 -0.048828011 > d3(9) d3(10) d3(11) d3(12) d3(13) d3(14) > -0.012357469 -0.042282285 0.010908775 -0.061393069 0.058384786 -0.036842060 > > $d4 > d4(0) d4(1) d4(2) d4(3) d4(4) d4(5) > d4(6) > 0.05419644 -0.23596007 0.84412709 0.06639234 0.02718491 0.15101792 > -0.18101631 >
Re: [R] How to reshape this data frame from long to wide ?
Not completely clear what you want (it does not appear to be a conventional reshape) but try this: > m <- matrix(c("A", "A", "B", "1", "2", "3"), 3, 2) > structure(do.call(cbind, lapply(tapply(m[,2], m[,1], c), ts)), tsp = NULL, > class = NULL) A B [1,] "1" "3" [2,] "2" NA On Sat, Feb 21, 2009 at 10:23 PM, Daren Tan wrote: > I tried cast and melt in reshape package, but still can't convert this data > frame m > > m > [,1] [,2] > [1,] "A" "1" > [2,] "A" "2" > [3,] "B" "3" > to this form. > > m1 > [,1] [,2] > [1,] "A" "B" > [2,] "1" "3" > [3,] "2" NA > Please help. > >[[alternative HTML version deleted]] > > __ > R-help@r-project.org mailing list > 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. > __ R-help@r-project.org mailing list 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.
[R] how to recover a list structure
I am experiencing some problems at working with lists at high level. In the following "coef" contains the original DWT coefficients organized in a list. Thorugh applying the following two commands: coef.abs <- lapply(unlist(coef,recursive=FALSE,use.names =TRUE),abs) coef.abs.sorted <- sort(unlist(coef.abs),decreasing=TRUE) I get vector "coef.abs.sorted" containing the coefficients absolute value ordered by decreasing magnitude, whose first 10 elements I set to zero after applying an iterative math formula (telling me how many coefficient values to spare). My problem is setting to zero the correspondent coefficients in the original "coef" list so as to reconstruct the signal from the depleted coefficients set, given that the above operations have erased the list structure. I can get the labels of the coefficients that have to be zeroed ... but the names of vector "coef.abs.sorted" elements differes from the original "coef" list elements as shown in the following: > names(coef.abs.sorted[1:10]) [1] "d6.d6(0)" "d5.d5(2)" "d5.d5(0)" "d4.d4(2)" "d5.d5(1)" "d4.d4(1)" "d4.d4(6)" "d4.d4(5)" "d4.d4(3)" [10] "d3.d3(12)" I wonder if there exists a high level command to set to zero respectively the coefficients in list "coef": - Level=6, Oscillation-Number=0 - Level=5, Oscillation-Number=0,1,2 - Level=4, Oscillation-Number=1,2,3,5.6 - Level=3, Oscillation-Number=12 > coef $d1 d1(0) d1(1) d1(2) d1(3) d1(4) d1(5) d1(6) d1(7) 1.075111e-02 -4.777441e-03 -2.775026e-03 2.011722e-03 -2.376611e-02 -1.574422e-03 -2.810372e-02 8.542393e-03 d1(8) d1(9)d1(10)d1(11)d1(12) d1(13)d1(14)d1(15) -1.754845e-02 -7.363856e-04 3.123922e-03 1.053743e-02 -1.014095e-02 1.059210e-02 -4.152311e-03 -7.320749e-04 d1(16)d1(17)d1(18)d1(19)d1(20) d1(21)d1(22)d1(23) -3.680064e-03 5.620248e-03 -2.251741e-03 -1.454674e-02 1.891275e-02 2.092009e-03 -1.286729e-02 9.037282e-03 d1(24)d1(25)d1(26)d1(27)d1(28) d1(29)d1(30)d1(31) -1.027137e-02 1.014230e-02 -2.175260e-03 -1.689383e-03 -2.016800e-03 3.984632e-03 -1.563781e-03 -3.801979e-03 d1(32)d1(33)d1(34)d1(35)d1(36) d1(37)d1(38)d1(39) 6.976144e-04 1.048419e-03 -1.319865e-02 -8.368520e-04 6.618890e-03 -5.835673e-03 8.721318e-03 2.018621e-02 d1(40)d1(41)d1(42)d1(43)d1(44) d1(45)d1(46)d1(47) -3.329934e-03 -4.823901e-03 2.957486e-02 -2.887600e-02 2.500062e-02 -2.207839e-02 1.516034e-02 4.328945e-05 d1(48)d1(49)d1(50)d1(51)d1(52) d1(53)d1(54)d1(55) 1.605634e-02 -2.360298e-02 7.875788e-03 -4.129065e-03 -1.014460e-02 -2.676679e-03 -9.953614e-03 -8.844977e-03 d1(56)d1(57)d1(58)d1(59) -2.141513e-02 -7.405836e-03 -5.078857e-03 8.480675e-04 $d2 d2(0)d2(1)d2(2)d2(3)d2(4)d2(5) d2(6)d2(7)d2(8) 0.033393680 0.015767682 -0.006008297 0.016899336 0.006661543 0.005297962 -0.005782728 -0.009358879 -0.012128449 d2(9) d2(10) d2(11) d2(12) d2(13) d2(14) d2(15) d2(16) d2(17) -0.014097771 0.003439004 -0.010154074 0.031203264 -0.057290615 0.004552577 0.061358997 -0.040730889 0.004064522 d2(18) d2(19) d2(20) d2(21) d2(22) d2(23) d2(24) d2(25) d2(26) -0.007208508 0.010426952 0.003855540 -0.011492580 0.035028591 0.029604957 -0.020069126 -0.018983905 -0.033710992 d2(27) d2(28) d2(29) 0.035848240 -0.014716622 0.008372893 $d3 d3(0)d3(1)d3(2)d3(3)d3(4)d3(5) d3(6)d3(7)d3(8) 0.020437693 -0.023658475 0.033448175 0.009689787 0.024750901 0.040022643 0.016532575 -0.035730498 -0.048828011 d3(9) d3(10) d3(11) d3(12) d3(13) d3(14) -0.012357469 -0.042282285 0.010908775 -0.061393069 0.058384786 -0.036842060 $d4 d4(0) d4(1) d4(2) d4(3) d4(4) d4(5) d4(6) 0.05419644 -0.23596007 0.84412709 0.06639234 0.02718491 0.15101792 -0.18101631 $d5 d5(0) d5(1) d5(2) 2.3514158 -0.4952582 -2.5589658 $d6 d6(0) -5.643476 > coef.abs.sorted d6.d6(0) d5.d5(2) d5.d5(0) d4.d4(2) d5.d5(1) d4.d4(1) d4.d4(6) d4.d4(5) d4.d4(3) 0.00e+00 0.00e+00 0.00e+00 0.00e+00 0.00e+00 0.00e+00 0.00e+00 0.00e+00 0.00e+00 d3.d3(12)d2.d2(15)d3.d3(13)d2.d2(13) d4.d4(0) d3.d3(8) d3.d3(10)d2.d2(16) d3.d3(5) 0.00e+00
Re: [R] How to reshape this data frame from long to wide ?
G'day Darren, On Sun, 22 Feb 2009 11:23:27 +0800 Daren Tan wrote: > I tried cast and melt in reshape package, but still can't convert > this data frame m > > m > [,1] [,2] > [1,] "A" "1" > [2,] "A" "2" > [3,] "B" "3" from the output, this does not look like a data frame to me but like a matrix. > to this form. > > m1 > [,1] [,2] > [1,] "A" "B" > [2,] "1" "3" > [3,] "2" NA I don't think that this result can be achieved. "A" and "B" would become the column names in the newly created data frame but not be entries in the data frame themselves. I presume you are looking for something like: R> m <- data.frame(var=c("A", "A", "B"), value=c("1", "2", "3")) R> m var value 1 A 1 2 A 2 3 B 3 R> m$id <- c(1,2,1) R> m var value id 1 A 1 1 2 A 2 2 3 B 3 1 R> library(reshape) R> ( res <- cast(m, id~...) ) id AB 1 1 13 2 2 2 R> res[, !(names(res) %in% "id")] AB 1 13 2 2 HTH. Cheers, Berwin === Full address = Berwin A TurlachTel.: +65 6516 4416 (secr) Dept of Statistics and Applied Probability+65 6516 6650 (self) Faculty of Science FAX : +65 6872 3919 National University of Singapore 6 Science Drive 2, Blk S16, Level 7 e-mail: sta...@nus.edu.sg Singapore 117546http://www.stat.nus.edu.sg/~statba __ R-help@r-project.org mailing list 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.
Re: [R] Convert a list to matrix
This should do what you want: > m <- list() > m[["A"]] <- 1 > m[["B"]] <- 2:3 > # get the maximum length > maxLen <- max(sapply(m, length)) > # create a new list with elements padded out with NAs > newM <- lapply(m, function(.ele){ + c(.ele, rep(NA, maxLen))[1:maxLen] + }) > do.call(rbind, newM) [,1] [,2] A1 NA B23 > On Sat, Feb 21, 2009 at 10:17 PM, Daren Tan wrote: > I would like to convert a list to matrix. This can be easily achieved via > do.call. The only problem is each element of the list has different length, > which causes the recycling of values. How can I have NA instead of recycled > values ? > > m <- list() > m[["A"]] <- 1 > m[["B"]] <- 2:3 > do.call(rbind, m) > [,1] [,2] > A11 > B23 > >[[alternative HTML version deleted]] > > __ > R-help@r-project.org mailing list > 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. > -- Jim Holtman Cincinnati, OH +1 513 646 9390 What is the problem that you are trying to solve? __ R-help@r-project.org mailing list 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.
[R] How to reshape this data frame from long to wide ?
I tried cast and melt in reshape package, but still can't convert this data frame m m [,1] [,2] [1,] "A" "1" [2,] "A" "2" [3,] "B" "3" to this form. m1 [,1] [,2] [1,] "A" "B" [2,] "1" "3" [3,] "2" NA Please help. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list 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.
[R] Convert a list to matrix
I would like to convert a list to matrix. This can be easily achieved via do.call. The only problem is each element of the list has different length, which causes the recycling of values. How can I have NA instead of recycled values ? m <- list() m[["A"]] <- 1 m[["B"]] <- 2:3 do.call(rbind, m) [,1] [,2] A11 B23 [[alternative HTML version deleted]] __ R-help@r-project.org mailing list 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.
[R] EM for Density Estimation
Hi, Which package would you recommend for an implementation of EM for density estimation (eg. mixture of Gaussian)? Thanks! Lars. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list 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.
Re: [R] Grouped bwplots?
On Fri, Feb 20, 2009 at 3:57 PM, Fredrik Karlsson wrote: > Dear list, > > I am sorry for asking you this, but I am trying to do again what I > thought I have done before, although this time it does not work. > > So, given the data set: > >> testdf <- data.frame(grfak=sample(c("One","Two"),size=100,replace=TRUE), >> panfak= sample(c("Yes","No"),size=100,replace=TRUE), xfak= >> sample(c("Yep","Nope"),size=100,replace=TRUE), d=rnorm(100)) > > I would like to do: > >> bwplot(d ~ xfak | panfak,data=testdf, groups= grfak) > > where the groups argument makes a difference (as in dividing each box > into two by group). > Is this possible to do? If so, what I am doing wrong? An alternative approach is to use ggplot2: install.packages("ggplot2") library(ggplot2) qplot(xfak, d, data=testdf, facets = ~panfak, fill = grfak, geom="boxplot") Hadley -- http://had.co.nz/ __ R-help@r-project.org mailing list 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.
Re: [R] NOT an R problem: cannot install packages from distant repository
It turned out that the directory where I had installed contributed packages (as defined by the environment variable R_LIB) was damaged, though it was not revealed by the usual disk maintenance tools (chkdsk, etc). I deleted it and created a new one, and everything looks fine now. Renaud 2009/2/20 Renaud Lancelot > Still searching what's going on In fact, I can download manually any > package and install it. The problem occurs when updating the html files. I > can reproduce it with this: > > > link.html.help() > Error in .readRDS(pfile) : unknown input format > > traceback() > 5: .readRDS(pfile) > 4: .packages(all.available = TRUE, lib.loc = lib) > 3: sort(.packages(all.available = TRUE, lib.loc = lib)) > 2: make.packages.html(lib.loc) > 1: link.html.help() > > Any hint ? > > Renaud > > > 2009/2/20 Renaud Lancelot > > I met today a computer crash and our maintenance officer had to reinstall >> some components of the OS (MS Windows XP Pro) as well as the Internet >> browser (among other things). Now, I cannot install packages from a distant >> repository: >> >> > utils:::menuInstallPkgs() >> Error in .readRDS(pfile) : unknown input format >> > traceback() >> 5: .readRDS(pfile) >> 4: .packages(all.available = TRUE) >> 3: .install.winbinary(pkgs = pkgs, lib = lib, contriburl = contriburl, >>method = method, available = available, destdir = destdir, >>installWithVers = installWithVers, dependencies = dependencies, >>...) >> 2: install.packages(NULL, .libPaths()[1], dependencies = NA, type = type) >> 1: utils:::menuInstallPkgs() >> > sessionInfo() >> R version 2.8.1 Patched (2009-02-17 r47956) >> i386-pc-mingw32 >> >> locale: >> >> LC_COLLATE=French_France.1252;LC_CTYPE=French_France.1252;LC_MONETARY=French_France.1252;LC_NUMERIC=C;LC_TIME=French_France.1252 >> >> attached base packages: >> [1] stats graphics grDevices utils datasets methods base >> >> other attached packages: >> [1] fortunes_1.3-6 >> >> loaded via a namespace (and not attached): >> [1] tools_2.8.1 >> >> * >> >> My browser and e-mail client are working, and I can display the list of >> packages when I run utils:::menuInstallPkgs(). >> >> Did anybody meet the same problem, and how can I solve it ? >> >> Renaud >> -- >> Renaud Lancelot >> EDEN Project, coordinator >> http://www.eden-fp6project.net/ >> >> UMR CIRAD-INRA "Contrôle des maladies animales exotiques et émergentes" >> Joint research unit "Control of emerging and exotic animal diseases" >> >> CIRAD >> Campus International de Baillarguet TA A-DIR / B >> F34398 Montpellier >> http://www.cirad.fr http://bluetongue.cirad.fr/ >> >> Tel. +33 4 67 59 37 17 - Fax +33 4 67 59 37 95 >> Secr. +33 4 67 59 37 37 - Cell. +33 6 77 52 08 69 >> >> > > > -- > Renaud Lancelot > EDEN Project, coordinator > http://www.eden-fp6project.net/ > > UMR CIRAD-INRA "Contrôle des maladies animales exotiques et émergentes" > Joint research unit "Control of emerging and exotic animal diseases" > > CIRAD > Campus International de Baillarguet TA A-DIR / B > F34398 Montpellier > http://www.cirad.fr http://bluetongue.cirad.fr/ > > Tel. +33 4 67 59 37 17 - Fax +33 4 67 59 37 95 > Secr. +33 4 67 59 37 37 - Cell. +33 6 77 52 08 69 > > -- Renaud Lancelot EDEN Project, coordinator http://www.eden-fp6project.net/ UMR CIRAD-INRA "Contrôle des maladies animales exotiques et émergentes" Joint research unit "Control of emerging and exotic animal diseases" CIRAD Campus International de Baillarguet TA A-DIR / B F34398 Montpellier http://www.cirad.fr http://bluetongue.cirad.fr/ Tel. +33 4 67 59 37 17 - Fax +33 4 67 59 37 95 Secr. +33 4 67 59 37 37 - Cell. +33 6 77 52 08 69 [[alternative HTML version deleted]] __ R-help@r-project.org mailing list 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.
Re: [R] variable/model selction (step/stepAIC) for biglm ?
Hi Chuck, Thanks for the guidelines. I was hoping someone in the group already experienced handling this type of task and have some handy code to share. I'll wait another day or two to see if someone responds with any more ideas or experience, and if nothing will come up, I might try my hand in your suggestion Cheers, Tal On Sat, Feb 21, 2009 at 8:09 PM, Charles C. Berry wrote: > On Sat, 21 Feb 2009, Tal Galili wrote: > >> Hello dear R mailing list members. >> >> I have recently became curious of the possibility applying model >> selection algorithms (even as simple as AIC) to regressions of large >> datasets. > > > Large in the sense of many observations, one assumes. > > But how large in terms of the number of variables?? > > If not too many variables, then you can form the regression sums of squares > for all 2^p combinations of regressors from a biglm() fit of all variables > as biglm provides coef() and vcov() methods. > > If it is large, then you most likely will need to do subsampling to reduce > the number to 'not too many' via lm() and friends then and apply the above > strategy. > > I searched as best as I could, but couldn't find any >> >> reference or wrapper for using step or stepAIC to packages such as >> biglm. > > > Surely any direct implementation of step() would be hopelessly long in > execution time. > > > HTH, > > Chuck > > >> >> Any ideas or directions of how to implement such a concept ? >> >> >> Best, >> Tal >> >> >> >> >> >> >> >> >> >> -- >> -- >> >> >> My contact information: >> Tal Galili >> Phone number: 972-50-3373767 >> FaceBook: Tal Galili >> My Blogs: >> www.talgalili.com >> www.biostatistics.co.il >> >> __ >> R-help@r-project.org mailing list >> 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. >> > > Charles C. Berry(858) 534-2098 >Dept of Family/Preventive > Medicine > E mailto:cbe...@tajo.ucsd.edu UC San Diego > http://famprevmed.ucsd.edu/faculty/cberry/ La Jolla, San Diego 92093-0901 > > > -- -- My contact information: Tal Galili Phone number: 972-50-3373767 FaceBook: Tal Galili My Blogs: www.talgalili.com www.biostatistics.co.il __ R-help@r-project.org mailing list 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.
Re: [R] variable/model selction (step/stepAIC) for biglm ?
On Sat, 21 Feb 2009, Tal Galili wrote: Hello dear R mailing list members. I have recently became curious of the possibility applying model selection algorithms (even as simple as AIC) to regressions of large datasets. Large in the sense of many observations, one assumes. But how large in terms of the number of variables?? If not too many variables, then you can form the regression sums of squares for all 2^p combinations of regressors from a biglm() fit of all variables as biglm provides coef() and vcov() methods. If it is large, then you most likely will need to do subsampling to reduce the number to 'not too many' via lm() and friends then and apply the above strategy. I searched as best as I could, but couldn't find any reference or wrapper for using step or stepAIC to packages such as biglm. Surely any direct implementation of step() would be hopelessly long in execution time. HTH, Chuck Any ideas or directions of how to implement such a concept ? Best, Tal -- -- My contact information: Tal Galili Phone number: 972-50-3373767 FaceBook: Tal Galili My Blogs: www.talgalili.com www.biostatistics.co.il __ R-help@r-project.org mailing list 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. Charles C. Berry(858) 534-2098 Dept of Family/Preventive Medicine E mailto:cbe...@tajo.ucsd.edu UC San Diego http://famprevmed.ucsd.edu/faculty/cberry/ La Jolla, San Diego 92093-0901 __ R-help@r-project.org mailing list 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.
Re: [R] difference between assignment syntax <- vs =
Patrick Burns wrote: 'The R Inferno' page 78 is one source you can look at. Patrick Burns wow .. nice! .. thanks for posting this reference. Esmail __ R-help@r-project.org mailing list 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.
Re: [R] Python and R
Different methods of performing least squares calculations in R are discussed in @Article{Rnews:Bates:2004, author = {Douglas Bates}, title= {Least Squares Calculations in {R}}, journal = {R News}, year = 2004, volume = 4, number = 1, pages= {17--20}, month= {June}, url = http, pdf = Rnews2004-1 } Some of the functions mentioned in that article have been modified. A more up-to-date version of the comparisons in that article is available as the Comparisons vignette in the Matrix package. On Fri, Feb 20, 2009 at 6:06 AM, Gabor Grothendieck wrote: > Note that using solve can be numerically unstable for certain problems. > > On Fri, Feb 20, 2009 at 6:50 AM, Kenn Konstabel wrote: >> Decyphering formulas seems to be the most time consuming part of lm: >> >> mylm1 <- function(formula, data) { >> # not perfect but works >> F <- model.frame(formula,data) >> y <- model.response(F) >> mt <- attr(F, "terms") >> x <- model.matrix(mt,F) >> coefs <- solve(crossprod(x), crossprod(x,y)) >> coefs >> } >> >> mylm2 <- function(x, y, intercept=TRUE) { >> if(!is.matrix(x)) x <- as.matrix(x) >> if(intercept) x <- cbind(1,x) >> if(!is.matrix(y)) y <- as.matrix(y) >> solve(crossprod(x), crossprod(x,y)) >> } >> >>> system.time(for(i in 1:1000) mylm2(EuStockMarkets[,-1], >>> EuStockMarkets[,"DAX"])) >>user system elapsed >>6.430.006.53 >>> system.time(for(i in 1:1000) mylm1(DAX~., EuStockMarkets)) >>user system elapsed >> 16.190.00 16.23 >>> system.time(for(i in 1:1000) lm(DAX~., EuStockMarkets)) >>user system elapsed >> 21.430.00 21.44 >> >> So if you need to save time, I'd suggest something close to mylm2 rather >> than mylm1. >> >> Kenn >> >> >> >> On Thu, Feb 19, 2009 at 8:04 PM, Gabor Grothendieck >> wrote: >>> >>> On Thu, Feb 19, 2009 at 8:30 AM, Esmail Bonakdarian >>> wrote: >>> >>> > Thanks for the suggestions, I'll have to see if I can figure out how to >>> > convert the relatively simple call to lm with an equation and the data >>> > file >>> > to the functions you mention (or if that's even feasible). >>> >>> X <- model.matrix(formula, data) >>> >>> will calculate the X matrix for you. >>> >>> > >>> > Not an expert in statistics myself, I am mostly concentrating on the >>> > programming aspects of R. Problem is that I suspect my colleagues who >>> > are providing some guidance with the stats end are not quite experts >>> > themselves, and certainly new to R. >>> > >>> > Cheers, >>> > >>> > Esmail >>> > >>> > Kenn Konstabel wrote: >>> >> >>> >> lm does lots of computations, some of which you may never need. If >>> >> speed >>> >> really matters, you might want to compute only those things you will >>> >> really >>> >> use. If you only need coefficients, then using %*%, solve and crossprod >>> >> will >>> >> be remarkably faster than lm >>> >> >>> >> # repeating someone else's example >>> >> # lm(DAX~., EuStockMarkets) >>> >> >>> >> y <- EuStockMarkets[,"DAX"] >>> >> x <- EuStockMarkets >>> >> x[,1]<-1 >>> >> colnames(x)[1] <- "Intercept" >>> >> >>> >> lm(y ~ x-1) >>> >> solve(crossprod(x), t(x))%*%y# probably this can be done more >>> >> efficiently >>> >> >>> >> # and a naive timing >>> >> >>> >> > system.time( for(i in 1:1000) lm(y ~ x-1)) >>> >> user system elapsed >>> >> 14.640.33 32.69 >>> >> > system.time(for(i in 1:1000) solve(crossprod(x), crossprod(x,y)) ) >>> >> user system elapsed >>> >> 0.360.000.36 >>> >> >>> >> >>> >> Also lsfit() is a bit quicker than lm or lm.fit. >>> >> >>> >> Regards, >>> >> Kenn >>> > >>> > __ >>> > R-help@r-project.org mailing list >>> > 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. >>> > >>> >>> __ >>> R-help@r-project.org mailing list >>> 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. >> >> > > __ > R-help@r-project.org mailing list > 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. > __ R-help@r-project.org mailing list 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.
Re: [R] Debian Power PC Compile R
On 20 February 2009 at 22:40, stephen sefick wrote: | I have the libx11-dev package installed through the aptitude | application. This is on 5.0 power pc. What can I provide to help | with this? I don't know where to begin. Sorry, but what exactly is your question? "Provide help with what?" If it is about building / using / extending / ... R on Debian/Ubuntu systems, come to the r-sig-debian list. Subscribe first before you can post. If you wonder about what you need to build R on Debian, the current Build-Depends entry in our debian/control file is Build-Depends: gcc (>= 4:4.1.0), g++ (>= 4:4.1.0), gfortran (>= 4:4.1.0), libblas-dev, liblapack-dev (>= 3.1.1), tcl8.5-dev, tk8.5-dev, bison, groff-base, libncurses5-dev, libreadline5-dev, debhelper (>= 5.0.0), texi2html, texinfo (>= 4.1-2), libbz2-dev, libpcre3-dev, xpdf-reader, zlib1g-dev, libpng12-dev, libjpeg62-dev, libx11-dev, libxt-dev, x11proto-core-dev, libpango1.0-dev, libcairo2-dev, libtiff4-dev, xvfb, xauth, xfonts-base, texlive-base, texlive-latex-base, texlive-generic-recommended, texlive-fonts-recommended, texlive-extra-utils, texlive-latex-recommended, texlive-latex-extra, texinfo, texi2html, openjdk-6-jdk [!arm !hppa !kfreebsd-i386 !kfreebsd-amd64 !hurd-i386] so you need a tad morethan libx11-dev. Dirk -- Three out of two people have difficulties with fractions. __ R-help@r-project.org mailing list 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.
[R] Bold fonts and greek characters in lattice plots
Hi, I am trying to "convince" lattice to use bold face in an expression containing Greek characters without any luck. Example code (without Greek characters, renders the expression in bold) library(lattice) data(Cars93,package=="MASS") splom(~Cars93[,5:8]|Origin,data=Cars93,panel=function(x,y,...) { panel.splom(x,y,...) dum<-format(cor(x,y,use="complete",method="kendal"),dig=2) panel.text(30,40,bquote(.(dum)),font=2) },pscales=0,col="gray" ) Example code (with Greek characters, no bold) data(Cars93,package=="MASS") splom(~Cars93[,5:8]|Origin,data=Cars93,panel=function(x,y,...) { panel.splom(x,y,...) dum<-format(cor(x,y,use="complete",method="kendal"),dig=2) panel.text(30,40,bquote(tau == .(dum)),font=2) },pscales=0,col="gray" ) Any suggestions? Christos Argyropoulos _ s. It's easy! aspx&mkt=en-us __ R-help@r-project.org mailing list 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.
Re: [R] difference between assignment syntax <- vs =
'The R Inferno' page 78 is one source you can look at. Patrick Burns patr...@burns-stat.com +44 (0)20 8525 0696 http://www.burns-stat.com (home of "The R Inferno" and "A Guide for the Unwilling S User") Thomas Mang wrote: Hi, Both operators <- and = can be used to make an assignment. My question is: Is there a semantic difference between these two? Some time ago, I remember I have read that because of some reason, one should be given preference over the other - but I cannot remember the source, nor the argument, nor which operator the preferred was. What is the present state ? Is still one version better than the other, or is it only a matter of taste what to use ? thanks Thomas __ R-help@r-project.org mailing list 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. __ R-help@r-project.org mailing list 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.
Re: [R] difference between assignment syntax <- vs =
Read the help page as to their differences: ?"<-" On Sat, Feb 21, 2009 at 7:30 AM, Thomas Mang wrote: > Hi, > > Both operators <- and = can be used to make an assignment. My question is: > Is there a semantic difference between these two? Some time ago, I remember > I have read that because of some reason, one should be given preference over > the other - but I cannot remember the source, nor the argument, nor which > operator the preferred was. > > What is the present state ? > Is still one version better than the other, or is it only a matter of taste > what to use ? > > thanks > Thomas > > __ > R-help@r-project.org mailing list > 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. > __ R-help@r-project.org mailing list 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.
[R] difference between assignment syntax <- vs =
Hi, Both operators <- and = can be used to make an assignment. My question is: Is there a semantic difference between these two? Some time ago, I remember I have read that because of some reason, one should be given preference over the other - but I cannot remember the source, nor the argument, nor which operator the preferred was. What is the present state ? Is still one version better than the other, or is it only a matter of taste what to use ? thanks Thomas __ R-help@r-project.org mailing list 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.
Re: [R] How to connect R and WinBUGS/OpenBUGS/LinBUGS in Linux in Feb. 2009
I just wanted to post in conclusion to this thread that I have had success running WinBUGS from R via R2WinBUGS, with the help of Gorjanc, Uwe, and Ben by email outside of this thread. I may have had a permissions problem, that was probably corrected by entering this in the terminal: m...@computer:~$ chmod -R u+w /home/me/.wine/drive_c/"Program Files"/WinBUGS14/ >From here, I opened R, entered library(R2WinBUGS), then ?bugs, then copy/pasted the example, and ran it exactly. My mistake that prevented it all from running was that I started R as sudo R, thinking that would give me more permissions, because I thought I was having permissions-oriented problems. But this is wrong. When I started R by merely entering R in the terminal, the example code ran perfectly. Success! Winepaths did not have to be specified because WinBUGS was installed in the usual place (c:/Program Files/WinBUGS14/"). Other people have emailed me, indicating that newer versions of WINE have not worked for them, so I am back with WINE 1.0. I hope this helps others trying to run WinBUGS on Linux. -- View this message in context: http://www.nabble.com/How-to-connect-R-and-WinBUGS-OpenBUGS-LinBUGS-in-Linux-in-Feb.-2009-tp22058716p22136577.html Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list 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.
Re: [R] importing data to SQLite database with sqldf
WARNING!!! Not a good idea to post code that deletes every object in your workspace! Other people may blindly copy and paste your code. I've added some verbiage to Example 12 on the home page: http://sqldf.googlecode.com that hopefully clarifies it a bit. On Sat, Feb 21, 2009 at 2:36 AM, Stephen Tucker wrote: > Thanks yet another time, Gabor - > I think I am slowly understanding - particularly I was confused by > persistence of connections. > > So starting with some parts of your example 12, > > ## ... deleted code which clears workspace ... > sqldf("attach 'mydb' as new") > irishead <- file("irishead.dat") > iristail <- file("iristail.dat") > > If I just wanted to merge the two files within SQL and return some part of > the result, I would do > > sqldf('select count(*) from (select * from irishead > union > select * from iristail)',dbname="mydb") > > and the tables exist in mydb only for the duration of the computation >> sqldf("select * from sqlite_master",dbname="mydb")$name > NULL > (but why is the size of mydb > 0 afterwards, if it contains no tables...?) > > ...is the above the same as > sqldf('select count(*) from (select * from irishead > union > select * from iristail)',dbname=tempfile()) > > except that I don't create 'mydb'? Yes. The third possibility is to omit dbname= entirely and then it uses a temporary "in memory" database. > > If I wanted to save the merged table (for use in a later session): > > sqldf('create table fulliris as select * from irishead > union > select * from iristail',dbname="mydb") > >> sqldf("select * from sqlite_master",dbname="mydb")$name > [1] fulltable > Levels: fulltable > > If I want copies of all three tables, > sqldf(dbname="mydb") > sqldf('create table fulltable as select * from irishead > union > select * from iristail') > sqldf() > >> sqldf("select * from sqlite_master",dbname="mydb")$name > [1] irishead iristail fulltable > Levels: fulltable irishead iristail > > ? ...I'll try to go figure a few more things out in the in the meantime (like > using sep="\t" ?) and using connections with sqldf(). > > But thanks for the help! > > Stephen > > - Original Message > From: Gabor Grothendieck > To: Stephen Tucker > Cc: R-help > Sent: Friday, February 20, 2009 5:22:09 AM > Subject: Re: [R] importing data to SQLite database with sqldf > > Have just added an example 12 on the home page: > > http://sqldf.googlecode.com > > that shows an example. Note use of notation > main.mytable to refer to an existing table in the > main database (as opposed to a data frame in R). > > On Thu, Feb 19, 2009 at 11:55 PM, Stephen Tucker wrote: >> Hi all, >> >> I am attempting to learn SQL through sqldf... >> >> One task I am particularly interested in is merging separate >> (presumably large) files into a single table without loading these >> files into R as an intermediate step (by loading them into SQLite and >> merging them there). >> >> Taking a step back, I've considered these alternatives: >> >> 1) I know if I use straight SQLite commands I might use the 'IMPORT' >> or 'INSERT INTO' command, which is not terribly flexible... (I can >> read large files line-by-line in Python and use the 'INSERT INTO' >> command, which is reasonably fast; I could do this in R as well but my >> experience with R's input/output is that it's much slower...? and >> sometimes setting up the table column definitions can be tedious if >> there are many variables). >> >> 2) dbWriteTable() with append=TRUE is very convenient except that it >> requires I load the data into R first... >> >> 3) sqldf's capability to put data directly into a database is >> something I'd like to work out. >> >> So in this case I have a series of tab-delimited text file with the >> first line being a header. >> >> For some reason I cannot seem to get it working. Combining examples 6 >> and 9 from the Google Code page (and R-help archives), I tried >> >> source("http://sqldf.googlecode.com/svn/trunk/R/sqldf.R";) >> (do I need it for SQLite?) >> ## >> sqldf("attach 'mydb.db' as new") >> f <- file("myexample.txt") >> attr(f,"file.format") <- list(header=TRUE,sep="\t") >> sqldf("create table myexample as select * from f", >> stringsAsFactors=FALSE, >> dbname="mydb.db") >> ## or >> f <- file(fi) >> sqldf("create table myexample as select * from f", >> stringsAsFactors=FALSE,file.format=list(header=TRUE,sep="\t"), >> dbname="mydb.db") >> ## >> sqldf("select * from myexample",dbname="mydb.db") >> gives me tables with 0 rows and 0 columns... >> >> So in any case I have a few questions: >> >> === 1 >> Would this be scalable to files with few GBs of data in them (I guess >> I am uncertain of the underlying mechanism for transporting data from >> the .txt file to the .db file... I see there is a call the >> dbWriteTable() internally in sqldf but through the connection)? And >> is there anything obviously doing wrong above? >> >> === 2 === >> Since I cannot 'append' rows to existing tables
[R] [R-pkgs] new RcmdrPlugin.survival package
Dear R users, I'd like to announce the RcmdrPlugin.survival package, which has just been made available on CRAN. The package provides an R Commander plug-in for the survival package, with dialogs for Cox models, parametric survival regression models, estimation of survival curves, and testing for differences in survival curves, along with data-management facilities and a variety of tests, diagnostics and graphs. All of this is tightly integrated with the standard R Commander menus. As does the Rcmdr, the package takes advantage of the R facilities for translating messages into other languages. It comes with a translation into Brazilian Portuguese, courtesy of Marilia Sá Carvalho of FIOCRUZ in Rio de Janeiro, whose suggestions during the development of the package were invaluable. Please contact me if you're interested in providing a translation into another language. Reports of bugs and other problems are welcome, as are suggestions for extending the package. Regards, John -- John Fox, Professor Department of Sociology McMaster University Hamilton, Ontario, Canada web: socserv.mcmaster.ca/jfox ___ R-packages mailing list r-packa...@r-project.org https://stat.ethz.ch/mailman/listinfo/r-packages __ R-help@r-project.org mailing list 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.
[R] variable/model selction (step/stepAIC) for biglm ?
Hello dear R mailing list members. I have recently became curious of the possibility applying model selection algorithms (even as simple as AIC) to regressions of large datasets. I searched as best as I could, but couldn't find any reference or wrapper for using step or stepAIC to packages such as biglm. Any ideas or directions of how to implement such a concept ? Best, Tal -- -- My contact information: Tal Galili Phone number: 972-50-3373767 FaceBook: Tal Galili My Blogs: www.talgalili.com www.biostatistics.co.il __ R-help@r-project.org mailing list 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.
[R] errors
Dear Sir/Madam, I was trying to upload my data in BRB array tools (R version 2.8.0). The following errors occured: Error: package/namespace load failed for 'affy' Error occured while executing the following R command: try(library(affy)). Could you please tell me what this means (and how to solve)? Thank you in advance, Henk-Marijn de Jonge De inhoud van dit bericht is vertrouwelijk en alleen bestemd voor de geadresseerde(n). Anderen dan de geadresseerde(n) mogen geen gebruik maken van dit bericht, het niet openbaar maken of op enige wijze verspreiden of vermenigvuldigen. Het UMCG kan niet aansprakelijk gesteld worden voor een incomplete aankomst of vertraging van dit verzonden bericht. The contents of this message are confidential and only intended for the eyes of the addressee(s). Others than the addressee(s) are not allowed to use this message, to make it public or to distribute or multiply this message in any way. The UMCG cannot be held responsible for incomplete reception or delay of this transferred message. [[alternative HTML version deleted]] __ R-help@r-project.org mailing list 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.
Re: [R] 24 hour axis /time plots
somewhereondearth wrote: Hi, i have a data with values against time and date. for plotting against date i used axis.Date() option. but I need a 24hr time plot as well. i have extracted and plotted, but not being able to put the time ticks ( eg. 6am, 7 am etc) on the x-axis. Im new to R, any help would be appreciated!! Hi somewhereondearth, I would suggest that you convert your hours into military time (-2359) and then plot the data without an X axis (xaxt="n"). Then add the X axis with the appropriate 12 hour labels: hr12labels<-c("12AM","2AM",...,"10PM") axis(1,at=seq(0,2400,by=200),labels=hr12labels) You can do this with date/time variables, but for a one-off, it may be easier to just define the labels you want yourself. Jim __ R-help@r-project.org mailing list 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.
Re: [R] [package-car:Anova] extracting residuals from Anova for Type II/III Repeated Measures ?
Dear John. Thank you so much for the patience and time, you have just answered my question! I was so used to the "mean ss" term from the aov table, that I didn't stop to look and realize that the values you added in the "Error SS" and "den Df" columns in the summary tables are actually the residuals sum of squares Errors and there respective degrees of freedom (and to that confusion also contributed, I guess, the fact that until your last e-mail I couldn't reproduce the aov results with Anova for me to experiment and compare). I hope that the clarification evoked by this discussion might have some use for future updates to the Anova (or aov) help files. And even if not - I personally feel to have gained a lot from this weeks correspondents. So again - thanks a lot John, for your patience and help. (p.s: I am adding " r-help@r-project.org" cc'd, so that other people encountering my question could benefit from your help as well) With regards, Tal On Sat, Feb 21, 2009 at 4:25 AM, John Fox wrote: > Dear Tal, > >> -Original Message- >> From: Tal Galili [mailto:tal.gal...@gmail.com] >> Sent: February-20-09 6:15 PM >> To: John Fox >> Subject: Re: [R] [package-car:Anova] extracting residuals from Anova for > Type >> II/III Repeated Measures ? >> >> Hello John, thanks for your reply and correction. >> >> I apologies for my crude mistake in applying the aov (now I have learned >> better). I hope to get a hold of "Statistical Models in S", but I don't >> predict it could easily happen in the near future. >> >> Also, I would be very happy if you could supply me with some more > directions >> as to how to obtain the "within" residuals (such as reported from the aov >> summary), since I am not sure how to proceed with that. > > I'm not sure what it is that you're looking for. First, I'm pretty sure that > you mean a residual sum of squares not residuals. But in any event, the sums > of squares reported by aov(), when you formulate the model correctly, are > exactly the same as those reported by Anova(). The "Error SS" given by > Anova() correspond to the "Residuals" sums of square given by aov(); these > are for the appropriate error term for testing each term. > > John > >> >> With regards, >> Tal >> >> >> >> >> >> >> >> >> >> On Sat, Feb 21, 2009 at 12:41 AM, John Fox wrote: >> >> >> Dear Tal, >> >> I didn't have time to look at all this yesterday. >> >> Since aov() doesn't do what I typically want to do, I guess I've not >> paid >> much attention to it recently. I can see, however, that you appear > to >> have >> specified the error strata incorrectly, since (given your desire to >> compare >> to Anova) the within-block factors are nested within blocks. > Something >> like >> >> > npk.aovE <- aov(value ~ N*P*K + Error(block/N*P*K), npk.long) >> >> should be closer to what you want, and in fact produces all of the > sums >> of >> squares, but doesn't put all of the error terms together with the >> corresponding terms; thus, you get, e.g., the test for N but not for > P >> and >> K, even though the SSs and error SSs for the latter are in the > table. >> By >> permuting N, P, and K, you can get the other F tests. I suspect that >> this >> has to do with the sequential approach taken by aov() but someone > else >> more >> familiar with how it works will have to fill in the details. I > wonder, >> though, whether you've read the sections in Statistical Models in S > and >> MASS >> referenced in the help file for aov. >> >> > summary(npk.aovE) >> >> >> Error: block >>Df Sum Sq Mean Sq F value Pr(>F) >> Residuals 5 153.147 30.629 >> >> >> Error: P >> >>Df Sum Sq Mean Sq >> >> P 1 16.803 16.803 >> >> Error: K >> >>Df Sum Sq Mean Sq >> >> K 1 190.40 190.40 >> >> >> Error: block:N >> >>Df Sum Sq Mean Sq F value Pr(>F) >> >> N 1 378.56 378.56 38.614 0.001577 ** >> Residuals 5 49.029.80 >> >> --- >> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 >> >> >> Error: block:P >> >>Df Sum Sq Mean Sq F value Pr(>F) >> >> Residuals 5 19.1317 3.8263 >> >> Error: block:K >> >>Df Sum Sq Mean Sq F value Pr(>F) >> >> Residuals 5 24.4933 4.8987 >> >> Error: P:K >> >> Df Sum Sq Mean Sq >> >> P:K 1 0.96333 0.96333 >> >> Error: block:N:P >> >>Df Sum Sq Mean Sq F value Pr(>F) >> >> N:P1 42.563 42.563 8.6888 0.03197 * >> Residuals 5 24.493 4.899 >> >> --- >> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 >> >> >> Error: block:N:K >> >>Df Sum Sq Mean Sq F value Pr(>F) >> >> N:K1 66.270 66.270 17.320 0.00881 ** >> Residuals 5 19.132 3.826 >> >> --- >> Sig
Re: [R] orthogonal/perpendicular distance from regression line
Hi Jeff, >> You could do this with [1], but if you are hoping to learn something >> statistical by doing this you should probably reconsider, because the >> regression line residuals are minimized along the dependent variable >> axis, It should be said that there are valid forms of linear regression where this is not true and it is the orthogonal residuals that are wanted. Perhaps the main case is where there are errors in x and y: indeed, this is how the "lines" (i.e. principal components) are fitted in principal component analysis. Regards, Mark. Jeff Newmiller wrote: > > On Fri, 20 Feb 2009, GAF wrote: > >> >> Hi there, >> I am trying to measure orthogonal/perpendicular distances from regression >> lines (i.e. the shortest distance from a given point to my regression >> line). >> As it sounds rather easy (basically just an orthogonal/perpendicular >> residual) I hoped that there was some function in R that can do that. All >> efforts so far remained unsuccessful, however. >> Does anybody know? > > You could do this with [1], but if you are hoping to learn something > statistical by doing this you should probably reconsider, because the > regression line residuals are minimized along the dependent variable axis, > and if the coordinate system is rotated (that is, you look at > deviation from the regression line to the points along a different > direction) then the residuals on that regression line will no longer be > minimized. > > [1] http://mathworld.wolfram.com/Point-LineDistance2-Dimensional.html > > --- > Jeff NewmillerThe . . Go > Live... > DCN:Basics: ##.#. ##.#. Live > Go... > Live: OO#.. Dead: OO#.. Playing > Research Engineer (Solar/BatteriesO.O#. #.O#. with > /Software/Embedded Controllers) .OO#. .OO#. > rocks...2k > > __ > R-help@r-project.org mailing list > 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. > > -- View this message in context: http://www.nabble.com/orthogonal-perpendicular-distance-from-regression-line-tp22123179p22134025.html Sent from the R help mailing list archive at Nabble.com. __ R-help@r-project.org mailing list 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.