... The short answer is that there's no such estimator in LingPipe 2.0. There are a couple of methods for pulling out n-gram collocations from a single...
Bob, I guess I should have given more context. I read a paper by David Carmel on lexical affinities to improve search precision. Lexical affinities are...
... basically collocations, ... LingPipe ... functionality I need. ... I'm afraid the short answer is still 'no'. Warning: The long answer is pretty long. I...
Hello all, I am writing the tutorials for the 2.0 release. I have one ready to go for text classification and am interested in feed back. I want to check that...
Greetings. I have modified the ne.Decoder class so that it writes in a file the scores of the HMM during tagging. I was planning to give the user a command...
Greetings. I was migrating from 1.0.7 to version 2.0 and I had the following problems: 1) I couldn't used the models created with the previous version. The ...
... When you say the "scores", what exactly do you mean - the per-token scores for the best path? ... These questions contain two parts that I'd like to tease...
... Thanks so much for the feedback. ... No. In fact, we're distributing exactly the same model files. I'm now wondering if we tested existing models. Was...
... I mean the scores per token for each possible pair of current_tag/previous_tag. ... I also considered extending the class and put it in the ne package in ...
... That'll be an awful lot of output if for even modest tag set sizes! But definitely more useful than just the deltas on the best path. For LingPipe 3.0,...
... Indeed, it is a lot, for 500K tokens the printed output is ~60MB. ... Looking forward to it, but can't wait until then! ... Well, for the time being the...
... <snipppity snip> ... file=/local/scratch/av308/temp/lingpipe/2.0Beta/in put/FBtemp.txtorg.xml.sax. ... despite ... The fix is on the website. There is also...
... Greetings. Yes, it works just fine now. Thanks. ... Thanks for this, but I am not sure how much of nice demo our problem would make. The general idea would...
Greetings. I was trying to play a bit with the sentence models and I had the following problem. I tried to use the HeuristicSentenceModel with the following...
The problem is that the HeuristicSentenceModel does not have a nullary (zero argument) constructor, as required by the command. If you want to use your own...
... Yes, thanks for the reply. A look at the code would have given me this answer without asking you! I just thought the HeuresticSentenceModel was ...
Hello, I am trying to use the Lingpipe training command. Initially, as a quick test I decided to use the sample.xml output file generated by the lingpipe demo....
... The problem is that LingPipe 1.7 output produces a lower case "type" tag, which breaks the parser--parser fixed in 2.0. Easiest workaround is to change the...
Hey I just wanted to tell you about a great company I found! I applyed for a lower mortgage rate and i only had to fill in a simple form. This month alone I...
Bored? Want to meet someone new? Chat with Real Women in YOUR AREA Just Call 1-800-335-9396 Only $0.99 a min with a major credit card. NO CONNECT FEES! No...
Hi people, I have trained the lingpipe to use speech data [as in the automatic speech recogniser transcripts of english broadcast news]. The named entity...
... The -1 occurs when the pronoun is unresolved to any entity. Can you post an example that you think should have worked? Something like "John Smith came by...
... the following sequence is from one of the development set. president <ENAMEX id="5" type="PER">clinton</ENAMEX> said at the geneva meeting that <ENAMEX...
... type=male_pronoun> ... anyone?? ... lingpipe for ... You appear to have changed the tag from "PERSON" to "PER" for the type. I don't think the matcher...
... Thanks a lot breck. That seems to have done the trick. It seems to get PERSON pretty well now. Does this mean, I must change ORG to ORGANISATION, DAT to...
... It wouldn't be a bad idea to use the "standard" types, and some coref may be affected but I can't recall right now. Pronouns are only going to happen...
... get ... sure. Also, am currently working with broadcast news. Will give the results to you as and when they are available. B PS: Thanks once again for your...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
