... In LingPipe 2.2, there's no way to serialize a dynamic classifier. This is because there's no way to serialize the dynamic language models on which the...
... We're running through a Yahoo newsgroup. You don't need to be a Yahoo member to sign up, but it does require registration simply to prevent spam. ... ...
... Thanks for the clear test case. I'm afraid it's our fault the doc wasn't clear enough, because we intended the dictionary training to supplement training...
For a short while now I've been using LingPipe for NE detection and a bit of co-referencing. I'm pretty new to this field and therefore I'm very actively...
d.fleurbaaij wrote: This is a very interesting post--Bob and I spent quite a bit of time talking about yesterday. Since it is long I will summarize my response...
Apologies for being such an idiot and wasting your time with my previous posts. It turned out I made a rookie mistake of assuming an iterator was returning ...
... We all make those mistakes--no problem--my personal worst being complaining that the Perl debugger would not accept input from the command line, posting to...
Hi, There is a coreference and NER mistake that Lingpipe coreference module is doing and that in my opinion can be easily corrected. In the bottom text both...
... As usual, details matter. So the existence of 'Michael Dell' really should not force all subsequent 'Dell's to be a PERSON. It gets worse in other domains...
I've noticed the coreference (gui demo) sometimes assigns tags of -1 to pronouns, and this value is never assigned to any tag other than pronouns (though...
... Good sleuthing. That's exactly what's going on. We should provide more doc for the demos. I'm hoping to roll out a whole range of demos soon that'll be...
We build a ton of demos that never see the light of day because they are not API worthy, or we don't have time to wrap them into a well documented tutorial. ...
Can someone describe the output of the sentiment tutorial. I understand it can determine whether a review is positive or negative or whether a sentence is...
... The tutorial is not working at that level of granularity. You are asking for phrase level sentiment which would require training data that was span ...
Hello all, I am presenting at the NY Java SIG (Special Interest Group) Tuesday, May 23rd. It's at Google. It will be an in depth presentation of how to do...
I am wondering if it would be helpful to have an evening or two a month where developers/students could come talk to Bob and I about LingPipe projects. I use...
At our Java SIG presentation this past Tuesday, we had several interesting questions about classifiers. One of the things people wanted to be able to do was...
... We will do better than that--we have a table at HLT/NAACL as will be giving out LingPipe 2.2.2 cds and will be happy to prototype your apps with LingPipe...
I will be at HLT/NAACL and will plan to stop by, but I don't have a laptop. If I bring some text files on a flash drive, would that work? -- Mary D. Taffet ...
Mary D. Taffet
mdtaffet@...
Jun 3, 2006 5:53 pm
293
... The most important thing to have (for ANYONE planning on having us help you get started) is data that you plan to process plus some idea of what you would...
I'm getting ready to upload medline baseline citations and annotations into a mysql database as suggested in the tutorial. I am using JDBC to implement a...
... Bob just re-ran results so I'll let him chime in with the numbers if he has them for bio entity recognition. ... Medline has lots of abstracts about things...
Here's the output from running the MEDLINE download script (as reported by an anonymous user): [java] Logged in to FTP Server. [java] Checking/Downloading...
Sorry this mail's getting so long, but it's a complex topic! ... Timing will vary based on the models you're using and how you configure the decoders. With...
I just got my lease and the directory name is: ftp.nlm.nih.gov/nlmdata/.medleasebaseline/gz Maybe they changed it recently? I can download from an ftp client...
I've finally tracked down my problem. We're not plugged in. Our IP address changed, so NLM would let us log in, but not see anything. That should be fixed...
I'm working on training the NE chunker to extract medical terms like diseases. The train method in the NETrain class takes an array of tokens and the...
... LingPipe expects the tags in BIO encoding (begin entity tag, continuing in-tag, and not in entity). It won't be able to do anything with DISEASE and...
Thanks Bob. So what should I specify as tags for each of the tokens in my example? Can you send me a sample training file? Thanks, -Ravi ... From: Bob...
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.
