... Thanks for pointing this out. I've patched Strings.decimalFormat(...) to return a sequence of "?"s instead of throwing an error in the case of overflow. ...
In our quest to get people comfortable working with LingPipe we are having our first Hobby Night in our Williamsburg Brooklyn office--the space is a very nice...
I've not yet implemented anything to save a trie dictionary. It "implements" the method in the Java API sense, so your code compiles. But as documented in ...
Hi, I have tried the Chinese segementation of LingPipe. But it seems that only for the same paragraph, the performance of the simplified Chinese is much worse...
... I should've made the style of Chinese more clear on the web page. What's running is the model trained on the Academia Sinica data. It's trained on...
... seems ... simplified ... tell why ... Thank you for your kind reply. I have read all the materials before. But after the compilation , I will only got the...
... Use the ant targets for training that you'll find in the Chinese word segmentation demo's build.xml file. For the Peking Uni data, use: ant train-pku05 You...
Hi, thank you for your kind reply. In fact ,I have read the materials before and I got the file: pk.out.knownWords and pk.out.segments. So I wonder how can I...
I'm responding to a message below about not being able to extract keywords from MEDLINE. This was broken in previous releases, but has been fixed. Are you...
Hi, I modified the simpleCoref to locate the male and female nouns: he, him, his, she, her, hers; ran it over the muc6 formal set; and used the muc software...
... Thanks. ... That makes sense, but it still only does mail pronouns. And couldn't the prefix Word boundary be brought out? And why the optional * on \\W? ...
Hi, I'm struggeling with moving from 2.2.1 to 2.3.0. I'm getting an array index out of bounds exception message of -1, when trying to the following code. The...
Hello all, Heng from NYU wanted go over cross document coreference as done by us (actually not part of the distribution) so I thought I would go ahead and make...
Breck, As there an ocean between us, coming over for the night isn't an option. But I'd very much like to follow the event! Is there any chance that you can ...
All, I wasn't really planning on much of a formal presentation. Not sure what it would achieve to record conversations around a monitor. The eve will force me...
Hi LP, One (or more) interesting problems that I wold like to solve, but I am not sure if LingPipe can help there. My domain are personal names, just like John...
... am not sure if LingPipe can help there. ... extent, company names. ... name (e.g. in some ZIP region). Cool problem. Did something like this a while back...
Hi Breck, Thanks for sparing here, If your model is this sensitive to noise we are going to have a problem in general. I am guessing that Bob will chime in and...
... I'm truly sorry about the big shift. I hope it's for the better after the transition, but I've been holding my breath waiting for messages like yours to...
A number of years back, during the launch of a monthly print publication (yes, that far back), I found the purchase of a set of CDs (yes, that far back)...
Hello, please forgive me it this is a silly question. Is it possible to evaluate the output of other (3rd-party) taggers, e.g. files annotated MUC-style with...
Florian Laws
florian@...
Aug 15, 2006 7:30 pm
343
... There's nothing built in, and it's not even clear how to do it from the doc. So I took this opportunity to add another section to the NE tutorial:...
Hmm. I guess Yahoo doesn't let me do attachments with mailing lists. Oh well. The last message had the files inlined (sorry about spacing), and here's the...
Up to now the cheapest point of entry for commercial use of LingPipe is $20,000–that is keeping us out of startups in garages..... I blogged about a new...
Hi, Having returned from vacation and almost dug my way out of my mail box, here are a couple of answers to your gracious ... The stats were computed using the...