R help - May 2012 - Topic Modeling- package lda

Hello,
I have been using 'lda' package for topic modeling and wish to predict
new topics using a fitted model by giving new dataset as input. While
doing this, I came across the function predictive.distribution(). But
I am not able to apply it on a new dataset as the input parameter asks
for document_sums which is an output of the result from fitting the
model. Is there any way in which I can use my existing model and apply
it on new data without creating a new model.

Any help will be appreciated.

Thanks in advance.
Ankit Sethi

Hi David,
I am sorry for cross posting. I will keep this in mind next time.
Regarding details, my doubt is regarding the documentation in the
package where, it speaks about applying the function to held-out words
but the input parameters don't take new datasets, rather they use the
output generated by the model. I wanted to get the doubt cleared and
now if my understanding of the function is right.

On Mon, May 7, 2012 at 8:13 PM, David Winsemius <dwinsemius at
comcast.net> wrote:
>
> On May 7, 2012, at 9:41 AM, ankit sethi wrote:
>
>> Hello,
>> I have been using 'lda' package for topic modeling and wish to
predict
>> new topics using a fitted model by giving new dataset as input. While
>> doing this, I came across the function predictive.distribution(). But
>> I am not able to apply it on a new dataset as the input parameter asks
>> for document_sums which is an output of the result from fitting the
>> model. Is there any way in which I can use my existing model and apply
>> it on new data without creating a new model.
>>
>
> This is identical to a very recent question on Stackoverflow. You are
> requested not to crosspost. You are also requested to provide considerably
> more detail. Please read the Posting Guide to get a better understanding of
> the "local culture".
>
> --
>
> David Winsemius, MD
> West Hartford, CT
>


-- 
Regards
Ankit Sethi