Irucka, You should cc R-help on all correspondence so that other readers of the list can follow the conversation.
(1) You say that you need 0s and 1s rather than TRUE/FALSE. Since a 0/1 matrix and TRUE/FALSE matrix behave exactly the same way in most applications, I'm not sure why it would matter which one you had. (2) My mistake. I had not noticed that you were eliminating the case where both the observed and the modeled were FALSE. I have modified my code to give the same results as your f2 function. compare.all <- sapply(modeldepth, function(md) mean((md==observeddepth)[md | observeddepth])) (3) My mistake again. There should be only single brackets, not double brackets. justbig <- modeldepth[compare.all > 0.7] Jean On Fri, Dec 21, 2012 at 11:27 AM, Irucka Embry <iruc...@mail2world.com>wrote: > Hi Jean Adams, how are you? > > I want to thank you for your response to my request for assistance. > > I received some assistance yesterday afternoon and I was able to update > the code which is posted here: > http://r.789695.n4.nabble.com/variable-names-in-numeric-list-and-Bayesian-inference-td4653674.html. > I posted the new code with some new questions that I have with regards to > the code that I have written. Can you look over that post and suggest any > code revisions for those aspects that do not work? Thank-you. > > The code that you suggested worked well overall, except for 3 aspects of > it: > > Here I actually needed the binary 0s and 1s rather than a TRUE/FALSE > logical matrix > > # a function to read in the data as a matrix of logicals > myreadfun <- function(file) { > as.matrix(read.ascii.grid(file)$data)!=0 > } > > Here I needed to calculate the f2 probability rather than the mean > > compare.all <- sapply(modeldepths, function(md) mean(md==observeddepth)) > > str(compare.all) > num [1:54] 0.936 0.94 0.944 0.944 0.945 ... > > Here most of the entries are greater than 0.7, but it should just be 31 of > the 54 that are greater than 0.7 > > > justbig <- modeldepths[[compare.all > 0.7]] > Error in modeldepths[[compare.all > 0.7]] : > recursive indexing failed at level 2 > > Once again, thank-you for your assistance. > > Irucka Embry > > > <-----Original Message-----> > >From: Adams, Jean [jvad...@usgs.gov] > >Sent: 12/21/2012 10:32:54 AM > >To: iruc...@mail2world.com > >Cc: r-help@r-project.org > >Subject: Re: [R] comparison of large data set > > > >Irucka, > > > > > >I did not test this code out on any data, but I think it will work. > > > > > >Jean > > > > > > > > > ># a function to read in the data as a matrix of logicals > >myreadfun <- function(file) { > >as.matrix(read.ascii.grid(file)$data)!=0 > >} > > > > > ># names of the 54 modeled depth files > >modfiles <- paste0("MaxFloodDepth_", 1:54, ".txt") > > > > > ># read in the observed and modeled depth files > ># observeddepth is a matrix > >observeddepth <- myreadfun("MaxFloodDepth_Observed.txt") > > > ># modeldepths is a list of matrices > >modeldepths <- lapply(filenames, myreadfun) > > > > > ># calculate the proportion of matrix elements that agree with the > observed file > ># the results is a vector with one number for each modeled depth matrix > >compare.all <- sapply(modeldepths, function(md) mean(md==observeddepth)) > > > > > ># select just those matrices that had a large proportion of agreements > ># justbig is a list of matrices > >justbig <- modeldepths[[compare.all > 0.7]] > > > >______________________________________________ > >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. > > _______________________________________________________________ > Get the Free email that has everyone talking at http://www.mail2world.com > Unlimited Email Storage POP3 Calendar SMS Translator Much More! > [[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.
