Named Entity Recognizers are a form of information extraction focusing precisely on named entities in order to classify them into specifically defined categories which may utilize entity linking. Annotation is a fundamental aspect of this classification. Quality measures often incorporate the use of precision, recall and F1 score (harmonic mean). Evaluations are also often compared against a gold standard: a benchmark that is available under reasonable conditions or the most accurate test possible without restrictions which is defined as the ground truth for the absolute state of information. The below highlight a few open source and commercial projects for NER. One can even utilize semantic web in form of a thesaurus server to incorporate SKOS schemes as a way of classification or annotation of terms in form of embedded URIs. One can view further examples from applications of PoolParty or Apache Stanbol.
OpeNER
Other Libraries for custom NER:
OpenNLP
UIMA
CORENLP
SPACY
NLTK
SyntaxNet & TensorFlow
DL4J
Apache Lucene
KEA
FastText
SpeedRead
Knowledge Population
Benchmarking
NER Survey
Google NLP API
Other Libraries for custom NER:
OpenNLP
UIMA
CORENLP
SPACY
NLTK
SyntaxNet & TensorFlow
DL4J
Apache Lucene
KEA
FastText
SpeedRead
Knowledge Population
Benchmarking
NER Survey
Google NLP API