search for: negll

Displaying 5 results from an estimated 5 matches for "negll".

Did you mean: negl
2007 Feb 17
1
Constraint maximum (likelihood) using nlm
...ng to find the maximum (likelihood) of a function. Therefore, I'm trying to minimize the negative likelihood function: # params: vector containing values of mu and sigma # params[1] - mu, params[2]- sigma # dat: matrix of data pairs y_i and s_i # dat[,1] - column of y_i , dat[,2] column of s_i negll <- function(params,dat,constant=0) { for(i in 1:length(dat[,1])) { llsum <- log( params[2]^2 + dat[i,2]^2) + (( dat[i,1] - params[1])^2/ (params[2]^2 + dat[i,2]^2)) } ll <- -0.5 * llsum + constant return(-ll) } Using (find data attached): data.osl <- read.table("osl.dat&...
2005 Sep 29
1
Error using a data frame as the "start" parameter in mle()
Dear R-Users, I am trying to use mle() to optimize two (or more) parameters, but I want to specify those parmeters in a data frame rather than having to spell them out separately in the "start" variable of mle(). My call is > mle(negll, start=list(aps=init), fixed=list(measphot=newphot, formod=formod, Nbands=Nbands), method="BFGS") where negll is a function I have written which uses the function predict.loess(). negll works fine when called directly. The parameter I am trying to optimize, "aps", is a data...
2011 Aug 17
2
An example of very slow computation
...t;-kvec[1] k2<-kvec[2] k3<-kvec[3] # MIN problem terbuthylazene disappearance z<-k1+k2+k3 y<-z*z-4*k1*k3 l1<-0.5*(-z+sqrt(y)) l2<-0.5*(-z-sqrt(y)) val<-100*(1-((k1+k2+l2)*exp(l2*t)-(k1+k2+l1)*exp(l1*t))/(l2-l1)) } # val should be a vector if t is a vector negll <- function(theta){ # non expm version JN 110731 pred<-Mpred(theta) sigma<-exp(theta[4]) -sum(dnorm(dat[,2],mean=pred,sd=sigma, log=TRUE)) } theta<-rep(-2,4) fand<-nlogL(theta) fsim<-negll(theta) cat("Check fn vals: expm =",fand," simple=",fsim,"...
2007 Dec 11
1
R computing speed
...trics, Academic Press. multilogit.c <- function(y, xi, xi.names = colnames(xi), c.base=1, rest=NULL, w = rep(1,nrow(y)), method='BFGS') { n.obs <- sum(w) xi<-cbind(1,xi) colnames(xi)[1]<-"Intercept" nx<-ncol(xi) ny<-ncol(y) beta<-numeric(nx*ny) negll<- function(beta,y,xi) { beta[rest]<-0 beta[(((c.base-1)*nx)+1):(c.base*nx)]<-0 lli <- y * (xi%*%matrix(beta,nx,ny) - log ( apply(exp( xi%*%matrix(beta,nx,ny)) ,1,sum ) ) ) lli<-lli*w -sum(lli) } pi<- apply((y*w),2,mean)/mean(w) ll0 <- (t(pi)%*%log(pi))*sum(w)...
2007 Jun 26
2
fisher information matrix
Hi All, a colleague wants to calculate the Fisher information matrix for a model he wrote (not in R). He can easily get the neg-log-likelihood and the best fit parameters at the minimum. He can also get negLLs for other parameter values too. Given these data, is there a way in R to calculate the Fisher information matrix? Best, Federico -- Federico C. F. Calboli Department of Epidemiology and Public Health Imperial College, St Mary's Campus Norfolk Place, London W2 1PG Tel +44 (0)20 7594 1602...