Re: [R] Error in optim while using fitdistr() function for estimationof parameters

From: Aswad Gurjar <aswadgurjar_at_gmail.com>
Date: Sun, 10 Feb 2008 12:56:26 +0530

Hello,

Thank You for response.Yes,It is discrete data.I have tried discrete model using goodfit() test.But that too is not working.Is there any other test for discrete data?
When I try goodfit() test for discrete distribution:

gf<-goodfit(x[[1]],type= "nbinomial",method= "MinChisq") where x[[1]] is my data with 9 bins.
x contains following data and interval is of 2800 seconds:

    V1
1 0
2 127
3 477
4 363
5 685
6 0
7 363
8 0
9 0
I get following errors:
Warning messages:

1: NaNs produced in: pnbinom(q, size, prob, lower.tail, log.p)
2: NaNs produced in: pnbinom(q, size, prob, lower.tail, log.p)
3: NaNs produced in: pnbinom(q, size, prob, lower.tail, log.p)
4: NaNs produced in: pnbinom(q, size, prob, lower.tail, log.p)
5: NaNs produced in: pnbinom(q, size, prob, lower.tail, log.p)
6: NaNs produced in: pnbinom(q, size, prob, lower.tail, log.p)
7: NaNs produced in: pnbinom(q, size, prob, lower.tail, log.p)


summary(gf)

         Goodness-of-fit test for nbinomial distribution

             X^2 df P(> X^2)
Pearson 5.539024 2 0.06269258
Warning message:
Chi-squared approximation may be incorrect in: summary.goodfit(gf)

When i use poisson distribution :
 gf<-goodfit(x[[1]],type= "poisson",method= "MinChisq")

I get following errors:
 Warning messages:
1: NA/Inf replaced by maximum positive value in: optimize(chi2, range(count))
2: NA/Inf replaced by maximum positive value in: optimize(chi2, range(count))
3: NA/Inf replaced by maximum positive value in: optimize(chi2, range(count))
4: NA/Inf replaced by maximum positive value in: optimize(chi2, range(count))
5: NA/Inf replaced by maximum positive value in: optimize(chi2, range(count))
6: NA/Inf replaced by maximum positive value in: optimize(chi2, range(count))
7: NA/Inf replaced by maximum positive value in: optimize(chi2, range(count))
8: NA/Inf replaced by maximum positive value in: optimize(chi2, range(count))
9: NA/Inf replaced by maximum positive value in: optimize(chi2, range(count))
10: NA/Inf replaced by maximum positive value in: optimize(chi2, range(count))
> summary(gf)

         Goodness-of-fit test for poisson distribution

                  X^2 df P(> X^2)
Pearson 4.137318e+138  3        0

Warning message:
Chi-squared approximation may be incorrect in: summary.goodfit(gf)

What should i do now?How can I remove these warnings and errors? Thank You.

Aswad
On 2/10/08, Bill.Venables_at_csiro.au <Bill.Venables_at_csiro.au> wrote:
>
> Chenge your model, most likely.
>
> You are telling us a) the response is discrete and b) there is a
> substantial probability that the value is zero. The gamma distributon
> is a model for a continuous response, your response is integer valued.
> Perhaps you should consider something the Poisson or Negative Binomial.
>
> Evem from the little you give us, the problem seems more statistical
> than computational to me.
>
>
> Bill Venables
> CSIRO Laboratories
> PO Box 120, Cleveland, 4163
> AUSTRALIA
> Office Phone (email preferred): +61 7 3826 7251
> Fax (if absolutely necessary): +61 7 3826 7304
> Mobile: +61 4 8819 4402
> Home Phone: +61 7 3286 7700
> mailto:Bill.Venables_at_csiro.au
> http://www.cmis.csiro.au/bill.venables/
>
> -----Original Message-----
> From: r-help-bounces_at_r-project.org [mailto:r-help-bounces_at_r-project.org]
> On Behalf Of Aswad Gurjar
> Sent: Sunday, 10 February 2008 4:17 PM
> To: r-help_at_r-project.org
> Subject: [R] Error in optim while using fitdistr() function for
> estimationof parameters
>
> Hello,
>
> I am trying to fit distribution for data consisting of 421 readings.It
> is
> basically no of requests arrived per minute.It contains many 0 entries
> as no
> of requests.When i use
> fd<-fitdistr(V2,"gamma")
> I get following error:
> Error in optim(x = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L,
> :
> initial value in 'vmmin' is not finite
>
> What should I do ? I need this function because I want to get value of
> estimated parameters.
> Is there any other function to perform such calculations?
> Please help me.
> Thank You.
>
> Code is as follows:
> x<-data.frame()
> N<-data.frame((read.table("cs2ip.txt")))
> attach(N)
> library(MASS)
> fd<-fitdistr(V2,"gamma")
>
> Aswad
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help_at_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.
>
>
>

        [[alternative HTML version deleted]]



R-help_at_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. Received on Sun 10 Feb 2008 - 07:29:17 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Sun 10 Feb 2008 - 11:30:13 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help. Please read the posting guide before posting to the list.

list of date sections of archive