I forgot to mention: You actually got correct results with using optim and `CG'. The warning messages were just telling you that there were some log(negative number) operations during the iterative process.
Ravi. ____________________________________________________________________ Ravi Varadhan, Ph.D. Assistant Professor, Division of Geriatric Medicine and Gerontology School of Medicine Johns Hopkins University Ph. (410) 502-2619 email: rvarad...@jhmi.edu ----- Original Message ----- From: arindam fadikar <arindam.fadi...@gmail.com> Date: Thursday, September 30, 2010 2:17 pm Subject: [R] how to avoid NaN in optim() To: r-help@r-project.org > hi , > > lik <- function(nO, nA, nB, nAB){ > > loglik <- function(par) > { > p=par[1] > q=par[2] > r <- 1 - p - q > > if (c(p,q,r) > rep(0,3) && c(p,q,r) < rep(1,3) ) > > { > -(2 * nO * log (r) + nA * log (p^2 + 2 * p * r) > + nB * log (q^2 + 2 * q * r) > + nAB * (log(2) +log(p) +log(q))) > } > else > NA > } > > loglik > > } > > > i want to maximize this likelihood function over the range (0,0,0) to > (1,1,1). so i tried > > optim ( c(0.3,0.3), lik ( 176,182 , 60 ,17) , method = "CG") > > and obtained the following : > > > optim(c(0.3,0.3), fn, method="CG") > $par > [1] 0.26444187 0.09316946 > > $value > [1] 492.5353 > > $counts > function gradient > 130 19 > > $convergence > [1] 0 > > $message > NULL > > Warning messages: > 1: In log(q^2 + 2 * q * r) : NaNs produced > 2: In log(q) : NaNs produced > 3: In log(q^2 + 2 * q * r) : NaNs produced > 4: In log(q) : NaNs produced > > > please help ... > > > -- > Arindam Fadikar > M.Stat > Indian Statistical Institute. > New Delhi, India > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help@r-project.org mailing list > > PLEASE do read the posting guide > 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.