Sorry -- should've calculated the AUC, not just looked at the results. Thanks for the detailed analysis. I didn't get your last e-mail, but yes, order of ties...
1241
Mike Ross
squawktopus
Oct 20, 2011 10:53 pm
I think you can get the best of both worlds in a single feature extractor 1) Do your onetime feature extraction and save results to a HashMap (serialized to...
1242
Bob Carpenter
colloquialdo...
Oct 21, 2011 3:48 am
Great idea -- that's a much better way to go than what I suggested. Can we fix the caching extractor so that it checks before putting so that you don't have to...
1243
Mike Ross
squawktopus
Oct 22, 2011 6:07 pm
Do you mean you want it to check a boolean parameter before putting? We could indeed to that, but its unnecessary and seems to go against the design concept...
1244
Bob Carpenter
colloquialdo...
Oct 26, 2011 7:58 pm
Amaç, Mike, and I have had some further back and forth about ROC curves, which I summarized on our blog because the discussion brought up some interesting ...
1245
Amac Herdagdelen
yalanciborsaci
Oct 31, 2011 2:50 pm
Interpolation is messy, I agree. For tie-breaking, I don't know how the implementation would be in a real area-under-the-curve computation. A quick&dirty...
1246
Yogesh
yogesh.pandi...
Nov 7, 2011 6:45 pm
How do I save the HashMap with both; category and filename along with the features. features = Map<String, Number> with filename (int) = Map<Integer,...
1247
Bob Carpenter
colloquialdo...
Nov 7, 2011 7:59 pm
I'm not clear where you're stuck because you outlined your own answer. The basic idea is like any other caching. As you perform expensive operations on an...
1248
Heather Dewey-Hagborg
hdeweyh
Nov 8, 2011 5:02 pm
I am working on a new sentence generating project that has a challenging requirement and I am looking for any advice or input into how it might be achieved. ...
1249
Bob Carpenter
colloquialdo...
Nov 8, 2011 6:33 pm
That sounds like it'd be an interesting app. The problem, as I'm sure you've realized, is that for a large vocabulary N, the number of pairs is N^2. Another...
1250
Mike Ross
squawktopus
Nov 8, 2011 9:33 pm
Adding to Bob's idea of mining for sentences with similar words and substituting...you could use WordNet to find similar words. Combined with POS tagging of...
1251
minh
minhtttran
Nov 9, 2011 3:49 pm
Good evening All, My name is Mi. I am a student at Georgia State in Atlanta, Georgia. I am new to LingPipe , and i would like to seek your help on how to do a...
1252
Nitish Ranjan
nitishranjan
Nov 9, 2011 4:02 pm
I would start with Lingpipe classification tutorial. I found the news group classification, easiest to follow. Once you have the code with you, start ...
1253
Heather Dewey-Hagborg
hdeweyh
Nov 9, 2011 4:39 pm
thanks for the ideas everyone - I'll let you know how it goes! -- Heather Dewey-Hagborg www.deweyhagborg.com 518-598-3775 [Non-text portions of this message...
1254
minh
minhtttran
Nov 9, 2011 8:25 pm
Hello Sinha, Could you please give me the link to the news group classification that you were mentioning? Thanks,...
1255
Nitish Ranjan
nitishranjan
Nov 9, 2011 8:27 pm
http://alias-i.com/lingpipe/demos/tutorial/classify/read-me.html Regards Nitish ... -- Have fun. [Non-text portions of this message have been removed]...
1256
Bob Carpenter
colloquialdo...
Nov 9, 2011 8:43 pm
You might want to look at our sentiment tutorial for something related to http://alias-i.com/lingpipe/demos/tutorial/sentiment/read-me.html To follow that, and...
1257
Dung Tran
dtran7@...
Nov 10, 2011 6:32 pm
Thank you very much for all of your help. May i ask you a basic question if you would not mind? I read and try to understand the sentimental tutorial but there...
1258
Nitish Ranjan
nitishranjan
Nov 10, 2011 6:38 pm
Minh, You will have to read some of these reviews yourself, this will serve as tagged data. Regards Nitish ... -- Have fun. [Non-text portions of this message...
1259
Alfian Akbar Gozali
panggil_aku_ian
Nov 10, 2011 11:57 pm
Hi all, Let me introduce myself, I am Alfian from Indonesia. I use lingpipe for my research in articles summarization for information retrieval. I think this...
1260
Bob Carpenter
colloquialdo...
Nov 11, 2011 12:59 am
1. Each corpus uses its own tagging scheme. For MedPost, it's described here: http://www.ncbi.nlm.nih.gov/staff/lsmith/papers/smith04a.pdf 2. That's an open...
1261
Rob
betterthanjimbo
Nov 11, 2011 2:50 am
Please remove me from your mailing list ... From: Bob Carpenter To: LingPipe@yahoogroups.com Sent: Thursday, November 10, 2011 7:59 PM Subject: Re: [LingPipe]...
1262
Bob Carpenter
colloquialdo...
Nov 11, 2011 5:37 am
I just took a look and don't think I can edit out members of the group. If you want to get out of the group, go to http://groups.yahoo.com/group/LingPipe/ and...
1263
Xibin Gao
xibingao
Nov 11, 2011 8:44 pm
Hi Bob, I am using different classifers (TradNaiveBayesClassifer, LogisticRegressionClassifer) for my text classification task and I would like to save the...
1264
Bob Carpenter
colloquialdo...
Nov 11, 2011 10:05 pm
If you compile a model to a stream, the result is a serialized Java object. So you can read it back in like any other serialized object using a...
1265
Bob Carpenter
colloquialdo...
Nov 11, 2011 10:08 pm
Seems to be homework day here at the mailing list. This is forwarded from mail sent to our help e-mail. Follow through the code in the clustering tutorial. 1....
1266
Alfian Gozali
panggil_aku_ian
Nov 12, 2011 2:49 am
Thank you for your reply :) It helped me much to do my research. But, still there is a question: In medpost and also genetag corpus, I found there are many...
1267
Bob Carpenter
colloquialdo...
Nov 14, 2011 3:29 am
I'm not an expert on the MedTag and GeneTag schemes. I don't know anything beyond the papers I sent. I think the "+" forms are either/or kinds of things where...
1268
pl_rudy
Nov 17, 2011 4:41 pm
Hi, I'm new to lingpipe, so bear with me and thanks for the help in advance. I'm trying to get word or token frequency in the document, I need them so that I...
1269
Bob Carpenter
colloquialdo...
Nov 17, 2011 5:02 pm
To get some background on how the LingPipe bits fit together, the best place for you to start would be the word counting tutorial: ...
