LingPipe 3.7.0 will generate a warning when compiling all jars because of an issue with non-ASCII chars in one of our java files. It works with Windows...
Hello, I was looking at NGramBoundaryLM and wondering why exactly those begin and end characters are needed and when you'd use NGramBoundaryLM instead of...
... The spelling corrector and HMMs wrap this all up the right way by default. Classifiers require you to specify which to use, and the decision is based on...
A user reported a problem with HMM decoding on a specific hardware/software configuration: Athlon64/Ubuntu 8.1 64-bit/JDK 1.6 An exception is thrown in the...
Bob - I recently got burned by http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6196102 going from Java 1.5 to Java 1.6 (1.6.0_11-b03, 32-bit, Linux on Intel...
... Thanks for the tip. I should've tried forcing interpreted mode to help debug. My 64-bit java doesn't have a client JVM, only a server one. I ran into some...
Out of curiosity, what happens when you hit this bug? Does the JVM just die (HotSpot error), or? Otis -- Sematext -- http://sematext.com/ -- Lucene - Solr -...
Yea, I guess they consider {@code <= Integer.MAX_VALUE} a rare case or something and its been fixed and regressed according to our cross JVM build experience....
... The problem was that a variable that's mathematically bounded to be between 0 and 1 with infinite precision wanders outside of that range with (some)...
Hi: I am a newbie of lingpipe. I read some tutorial on the langpipe website but many NLP terms frightened me. The classification samples included in the...
... There are also classifiers used in the language ID, word sense disambiguation, sentiment, and logistic regression tutorials. The word sense disambiguation...
... I am planning to crawl universities, research labs and job websites to get many researchers' profile information and then provide a search interface for...
... There's been some serious work on more or less exactly this problem of classifying university web pages into types. Mark Craven created a data set back in...
Hi: Many thanks for your help. Web-KB is a great corpus for me. As us said, logistic regression classifier is much harder to train. Maybe I have to dig into...
Hello I am trying out the Sentiment Analysis tutorial (PolarityHierarchical.java). I want to dump the classifier as a file, so that I donot have to run through...
... The answer is that the result is an LMClassifier, not a DynamicLMClassifier; the result of compilation is a more compact and efficient implementation, but...
Hi, bob: I am trying to use NaiveBayesClassifier to classify text documents. Here is my code: String regex = "[a-zA-Z]+"; RegExTokenizerFactory tf = new...
Thanks for the great bug report. Your diagnosis is sport on. Our RegExTokenizerFactory's serialization is broken, because I forgot the write-replace. I just...
Hi, bob: I have add the fix as u said, but it doesn't work for me now. And the same exception occurs. I have no idea about that. I dig into the code and find...
I'm afraid to say that you're absolutely right again. I fixed the serialization problem for RegExTokenizerFactory, but that wasn't at the root of the problem...
Hi,when I run the demo of Chinese words segmentation with Peking University corpus,OutOfMemoryError happened as follows: run-pk: [java] CHINESE TOKENS [java]...
Hi, I was using a custom LineParser to parse a file, when I ran into the following problem: ## Exception in thread "main" java.lang.NullPointerException at...
... I'm afraid that'll depend on the JVM. Using the 64-bit 1.6 JVM on Windows, it runs in 512M but runs out of memory with only 256M. It'll require less on...
Thanks so much for the detailed bug report and diagnosis (which I don't repeat here). You're absolutely right on both counts. 1. Here's what I found in the...
Thanks for the note and the patch. We really want the tutorials to work as written, so any suggestions for patches are greatly appreciated. I'm in the final...
LingPipe 3.8.0 is now available from the site: http://alias-i.com/lingpipe This is a big release — the details are available on the home page linked above. ...
I accidentally left an old description in the documentation on the LingPipe home page. We have not in fact deprecated util.FeatureExtractor as we said on the...
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.
