That's called "interval censoring". The "likelihood" for such cases is the probability of what was observed, i.e., (pgamma3(10, ..., - pgamma3(0, ...))*... . For this kind of problem, I have in the past written a function to compute the log(likelihood) and then passed that function to "optim".

hope this helps. spencer graves

Mårten Bjellerup wrote:

I have 'grouped' data, that is in the form of:

Interval        Median
0-9.9%:        -25
10-19.9%:     0
20-29.9%:     3
30-39.9%:     10
40-49.9%:     50
50-59.9%:     200
et cetera

and want to fit a three parameter gamma distribution. Does anyone know of an existing routine for doing this (or something similar)? Any help or comment is much appreciated.

Regards,

Mårten

Mårten Bjellerup
Doctoral Student in Economics
School of Management and Economics
Växjö University
SE-351 95  Växjö
Sweden

Tel: +46 470 708410 Fax: +46 470 82478 Mobile: +46 70 969 88 88 Mail: [EMAIL PROTECTED] Web: http://www.ehv.vxu.se
-------------------------------------
"Forecasting is like trying to
drive a car blindfolded and
following directions given by a person who is looking
out of the back window"


______________________________________________
[EMAIL PROTECTED] mailing list
https://www.stat.math.ethz.ch/mailman/listinfo/r-help



______________________________________________ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help

Reply via email to