Search Mailing List Archives
[java-nlp-user] 答复: NER - CRFClassifier Classes
dgerber at informatik.uni-leipzig.de
Thu Apr 7 06:29:41 PDT 2011
Thanks for your answer.
I've tried to train the classifier with my annotated file (as described at http://nlp.stanford.edu/software/crf-faq.shtml#b). This seems to work well with the Stanford NER package. However I'm using the (most-recent) Stanford-CoreNLP in my project and with this, it's not possible to train the classifier nor to load the trained one (from NER package). I'm getting this error message:
Exception in thread "main" java.lang.RuntimeException: java.io.InvalidClassException: edu.stanford.nlp.util.Index; local class incompatible: stream classdesc serialVersionUID = 5398562825928375260, local class serialVersionUID = 533612384823468898
I know multiple people dealt with this error before, but I found no real answer to my problem. Is there any way to train my model with this release?
One more thing: Is it somehow possible to add the trained DATE class to let's say the "conll.closed.iob2.crf"-classifier? I would really like to avoid recreating it or using different classifiers in my code?
P.S.: I would be happy to share my results with this list.
On 28.03.2011, at 20:29, Dakan Wang wrote:
> 1. For a detailed description of the data, check
> The MISC means any type other than PERSON, ORGANIZATION, LOCATION.
> 2. If you have a new entity type, you should train your own classifier and
> provide enough entities for the new type so that the classifier can learn
> from your data. For how to train your own classifier, check
> 发件人: java-nlp-user-bounces at lists.stanford.edu
> [mailto:java-nlp-user-bounces at lists.stanford.edu] 代表 Gerber Daniel
> 发送时间: 2011年3月28日 2:43
> 收件人: java-nlp-user at mailman.stanford.edu
> 主题: [java-nlp-user] NER - CRFClassifier Classes
> Hi all,
> I've integrated the AbstractSequenceClassifier into my software research
> project. Therefore I loaded the "conll.closed.iob2.crf.ser.gz"
> CRFClassifier. The classes supported by this classifier seem to be PERSON,
> ORGANIZATION, LOCATION and MISC. So my first question is:
> 1. What exactly does MISC stand for?
> And the second question is:
> 2. Is there any way (preferably, without to train my own classifier) to also
> support the class DATE?
> Kind regards,
> java-nlp-user mailing list
> java-nlp-user at lists.stanford.edu
More information about the java-nlp-user