I got the same error message in R 1.8.1 and S-Plus 6.1. Then I
read the error message: "'start' must be a named list". Then I read
help("fitdistr") and learned that "fitdistr" required 3 arguments, the
third of which was "start: A named list giving the parameters to be
optimized with initial values. This can be omitted for some of the
named distributions (see Details)." The normal distribution was NOT one
of the named distributions. When I supplied a named list for "start",
it worked:
> fitdistr(x, "normal", list(mean=0, sd=1))
mean sd
2.97093013 0.88852969
(0.08885297) (0.06281083)
hope this helps. spencer graves
Rau, Roland wrote:
-----Original Message-----
From: DivineSAAM at aol.com [SMTP:DivineSAAM at aol.com]
Sent: Sunday, February 15, 2004 10:24 PM
To: edwardweisun at hotmail.com; r-help at stat.math.ethz.ch
Subject: Re: [R] Maximum likelihood estimation in R
Hello,
Use
x=rnorm(100, mean=3, sd=1)
library(MASS)
fitdistr(x, "normal")
mean sd
2.93666631 0.99673982
(0.09967398) (0.07048015)
For me, this example does not work. As it looks like copy+paste, I
guess this something platform dependent? At least I have to specify some
starting values as shown below. But maybe I did something wrong?
Best,
Roland
(I am using at the moment R 1.8.1 on WinNT)
x=rnorm(100, mean=3, sd=1)
library(MASS)
fitdistr(x, "normal")
Error in fitdistr(x, "normal") : 'start' must be a named list
fitdistr(x, "normal", start=list(mean=1, sd=0.4))
mean sd
3.03167815 1.06637851
(0.10663786) (0.07539617)
+++++
This mail has been sent through the MPI for Demographic Research. Should you receive a mail that is apparently from a MPI user without this text displayed, then the address has most likely been faked. If you are uncertain about the validity of this message, please check the mail header or ask your system administrator for assistance.