Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.

# Re: st: Fit chi2 as in gammafit

 From Nick Cox <[email protected]> To [email protected] Subject Re: st: Fit chi2 as in gammafit Date Tue, 30 Aug 2011 21:27:35 +0100

```It doesn't worry me personally. I've never wanted to compare a
chi-square distribution with data.

On Tue, Aug 30, 2011 at 9:24 PM, Tirthankar Chakravarty
<[email protected]> wrote:
> Fair point Nick - one which I neglected to take into account.
>
> T
>
> On Tue, Aug 30, 2011 at 1:19 PM, Nick Cox <[email protected]> wrote:
>> Ho hum. A conservative might want the free parameter to be integer-valued.
>>
>> Nick
>>
>> On Tue, Aug 30, 2011 at 9:11 PM, Tirthankar Chakravarty
>> <[email protected]> wrote:
>>> You could do this pretty easily using -ml-, allowing additionally the
>>> possibility of including covariates to parameterise the mean
>>> (exponential) of the distribution.
>>>
>>> /*******************************************/
>>> // chi2 regression
>>> webuse dollhill3, clear
>>> poisson deaths smokes i.agecat
>>>
>>> cap prog drop chi2reg
>>> program chi2reg
>>>        version 11
>>>        args lnf theta
>>>        qui replace `lnf' = -0.5*exp(`theta')*log(2) ///
>>>                        -lngamma(exp(`theta')/2) + ///
>>>                        (exp(`theta')/2 - 1)*log(\$ML_y1) ///
>>>                        -\$ML_y1/2
>>> end
>>>
>>> ml model lf chi2reg (deaths = smokes i.agecat)
>>> ml maximize
>>> /*******************************************/
>>>
>>> T
>>>
>>>
>>> On Tue, Aug 30, 2011 at 12:53 PM, Maarten Buis <[email protected]> wrote:
>>>> Acutually you can use -gammafit- to estimate a chi square distribution
>>>> by constraining [beta]_b[_cons] to be 2. The estimated degrees of
>>>> freedom are than 2*[alpha]_b[_cons]:
>>>>
>>>> *------------ begin example -------------
>>>> drop _all
>>>> set obs  100
>>>> gen chi2 = rchi2(3)
>>>> constraint 1 [beta]_b[_cons]=2
>>>> gammafit chi2, constraint(1)
>>>>
>>>> // display estimated degrees of freedom:
>>>> lincom [alpha]_b[_cons]*2
>>>> *------------- end example ---------------
>>>> (For more on examples I sent to the Statalist see:
>>>> http://www.maartenbuis.nl/example_faq )
>>>>
>>>> Hope this helps,
>>>> Maarten
>>>>
>>>> On Tue, Aug 30, 2011 at 8:43 PM, Nick Cox <[email protected]> wrote:
>>>>> -gammafit- (SSC) is also to be attributed to Stephen Jenkins.
>>>>>
>>>>> I think the short answer is that nothing is canned and public, but
>>>>> existing code like -gammafit- would get you most of the way. Fitting a
>>>>> more general distribution has the merit that it tells you something of
>>>>> what ways the distribution departs from the specific case.
>>>>>
>>>>> See also official -pchi- and -qchi-.
>>>>>
>>>>> Nick
>>>>>
>>>>> On Tue, Aug 30, 2011 at 7:30 PM, Laurie Molina <[email protected]> wrote:
>>>>>> Hi,
>>>>>> I would like to fit a chi2 probability distribution to my data, in the
>>>>>> same way as the gammafit command (by Nick Cox) does it.
>>>>>> Is there any way to perform such a task?
>>>>>>
>>
>> *
>> *   For searches and help try:
>> *   http://www.stata.com/help.cgi?search
>> *   http://www.stata.com/support/statalist/faq
>> *   http://www.ats.ucla.edu/stat/stata/
>>
>
>
>
> --
> Tirthankar Chakravarty
> [email protected]
> [email protected]
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
```