Hi Caio, I just replied to a fellow who needed to run a series of regressions changing the dependent variables for the same period. I believe your solution might be similar to that one, just making the window dynamic instead. I used a lapply to store regressions and sapplys to extract stats from the lm()s. Please check the attachment and hope it´s useful.
Filipe Botelho -----Mensagem original----- De: r-help-boun...@r-project.org [mailto:r-help-boun...@r-project.org] Em nome de Caio Ramos Casimiro Enviada em: terça-feira, 31 de maio de 2011 12:23 Para: r-help@r-project.org Assunto: [R] Using lm() combined with sapply Dear list, I spent a lot of time looking for the solution for this problem, but with no success. It's the first time I came to the list to ask for help, so I'm sorry if I break some protocol. Well, I'm trying to make a serie of regressions in different periods, actually, in growing periods. In other words, I start by make a regression in a period from 1 to t, then I make another regression from 1 to t+1, and so on until t+1 < N. A loop version for this is: a <- rep(1,100)+rnorm(100) b <- rep(2,100)+rnorm(100) c <- rep(1.5,100)+rnorm(100) n.start <- 50 coefs <- array(NA,c(51,4)) for (i in (n.start:100)) coefs[i-n.start+1,] <- coef(lm(a ~ b + c, subset=1:i)) I want to avoid loops, so I was trying to make something like: regride <- function(formula,dat,sub,w) { window <- c(1:sub) w[1:sub-1] <- w[1:sub-1]*beta^sub coefs(lm(formula=formula,data=dat,subset=window,weights=w)) } coefs <- sapply(n_start:dim(ind)[1],regride,formula,dat,x) But it's not working. Best regards, Caio Casimiro -- Caio Ramos Casimiro Software Developer +55 11 9773 2809 [[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. "This message and its attachments may contain confidential and/or privileged information. If you are not the addressee, please, advise the sender immediately by replying to the e-mail and delete this message." "Este mensaje y sus anexos pueden contener información confidencial o privilegiada. Si ha recibido este e-mail por error por favor bórrelo y envíe un mensaje al remitente." "Esta mensagem e seus anexos podem conter informação confidencial ou privilegiada. Caso não seja o destinatário, solicitamos a imediata notificação ao remetente e exclusão da mensagem."
--- Begin Message ---Hi Raphael, using your data as is, and if I understood what you need > birds Square Sp1 Sp2 Sp3 Sp4 Spn Natprop Effort [1,] 1 1 0 1 1 0 0.5 10 [2,] 2 1 0 1 1 0 0.6 20 [3,] 3 1 1 0 1 0 0.8 23 [4,] 4 1 0 1 0 0 0.8 50 [5,] 5 0 1 0 1 1 0.9 30 I store all results from regressions in a list to then extract what I want bird_regressions <- lapply(1:5, function(i) lm(birds[,i+1] ~ birds[, "Natprop"] + birds[, "Effort"])) # Will just loop through Sp1, Sp2... Spn keeping "Natprop" and "Effort" fixed Now retrieving some important stats r_sq <- sapply(1:length(bird_regressions),function(i) summary(bird_regressions[[i]])$r.squared) betas <- sapply(1:length(bird_regressions),function(i) summary(bird_regressions[[i]])$coefficients[2,1]) tstat <- sapply(1:length(bird_regressions),function(i) summary(bird_regressions[[i]])$coefficients[2,3]) Now with cbind(r_sq, betas, tstat) You will have a matrix summarizing what I believe you need from Sp1, Sp2... Spn. Hope this was helpful. -----Mensagem original----- De: r-help-boun...@r-project.org [mailto:r-help-boun...@r-project.org] Em nome de Rafael Xavier de Camargo Enviada em: segunda-feira, 23 de maio de 2011 13:05 Para: r-help@r-project.org Assunto: [R] Linear regression - several response variables vs few indvariables Hi all, I need to run several simple linear regressions at once, using the following data. Response variables: Bird species (sp 1, sp2, sp3...spn). Independent variable: Natprop - proportion of natural area. covarate: Effort = hours). One single linear regression would be: lmSp1 <- lm(sp1~ natprop + effort). However, I need to run this linear regression for all bird species that I have individually (n = 163). I would like to do it at once and store the coefficients in a single data frame. Is that possible? Table that I have: Square Sp1 Sp2 Sp3 Sp4 Spn Natprop Effort 1 1 0 1 1 0 0.5 10 2 1 0 1 1 0 0.6 20 3 1 1 0 1 0 0.8 23 4 1 0 1 0 0 0.8 50 n 0 1 0 1 1 0.9 30 Thanks in advance. Rafael. ********************************************** RAFAEL CAMARGO Postgraduate Student Biology Department of University of Ottawa 30 Marie Curie, room # 351 Ottawa, ON, CANADA Tel: +1 (613) 562-5800 ext. 6366 Cel: +1 (613) 869-3772 e-mail: rcama...@uottawa.ca rafael.x.cama...@gmail.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.
--- End Message ---
______________________________________________ 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.
