On 06/17/2014 04:00 AM, Pathima Nusrath Hameed wrote:
I am trying to aggregate OpenNlpTextAnalyzer.pear and/or chunker.xml with
my own descriptors.
But it produces errors. Also it does not support the component descriptor
editor for change configurations. Does anybody know how to use opennlp
descriptors in uima? I am glad if you can help me.
To really make use of the UIMA integration you have to construct your
own AAE out of them
customized for your analysis needs.
E.g. for named entity detection:
Sentence Detector -> Simple Tokenizer -> Person Name Finder -> Location
Name Finder
In the beginning and the end of the pipeline you might need custom AEs
to load and store the
data.
UIMA takes some effort to get running, but when it works it is usually
pretty reliable.
HTH,
Jörn
