How many text titles do you have? You may have to create your own training set and train your own chunking model.
On Tue, Jul 17, 2012 at 7:24 PM, Sheng Guo <[email protected]> wrote: > > Hi Adam, > I guess I may need a better NER other than a chunker.I don't know if the > result will be significantly different when the title is short or sometimes > only a long phrase. Does current chunker or NER in OpenNLP handle these kind > of title-style text? > Thanks, > Sheng > >> From: [email protected] >> Date: Tue, 17 Jul 2012 17:28:14 -0400 >> Subject: Re: Any good way to chunk the news title? >> To: [email protected] >> >> Hi Sheng, >> >> As you mentioned, sometimes titles are just short phrases. Given this, the >> title might usually just be one chunk. What sort of results are you looking >> for? >> >> Adam >> >> On Tue, Jul 17, 2012 at 2:28 PM, Sheng Guo <[email protected]> wrote: >> >> > >> > Hi, >> > Recently I am trying the openNLP for chunking news title. News titles look >> > very different from normal sentence, sometimes it is just a short phrase. >> > I tried the current openNLP chunker with the default models, the result is >> > not that good. >> > Can anyone help me with this problem? >> > (1) Anyone has good chunker usage sample code? just make sure I am doing >> > it in the correct way.(2) Anyway to improve the short title chunking for a >> > specific domain? >> > Thanks! >> > Sheng >> >> >> >> >> -- >> *Adam Goodkind * >> *w* adamgoodkind.com <http://www.adamgoodkind.com> >> *t* @adamgreatkind <https://twitter.com/#%21/adamgreatkind> > -- Lance Norskog [email protected]
