On 2/1/11 5:52 AM, Sampath Kumar wrote:
Hi,
I am having issues with sentence detectors. I would like to not split at
abbreviated names and words. For e.g. If the sentence is "John D. and Jane D.
presented at the Intl. Conf. On XYZ", the default implementation splits after
the names and also after "Intl" and "Conf.". I tried to train the model by
giving similar sentences. Added about 1000 of the same kind. But I still get the
same result. Could you please help me in resolving this.
If you have 1000 of the same kind it should be able to detect these
cases I think,
maybe something was wrong with you training data?
You must place every sentence in one line.
The English model on the website is trained with around 50k sentences.
Jörn
