CRFChunker: A Tool for chunking sentences into chunks
CRFChunker is a Conditional Random Field (CRF) based chunker developed using CoNLL-2000 shared task data. We evaluated its performance against the gold chunks of interpretable Semantic Textual Similarity (iSTS) 2015 data: the training and test data sets each consisting of 375 pairs of Images annotation data and 378 pairs of Headlines texts. For the training set, the CRFChunker yielded accuracies of 86.20% and 68.34% at chunk level and sentence level respectively. Similarly, the chunker accuracies were 86.81% and 69% for the test set at chunk level and sentence level respectively.
The tool is also available as a part of SemAligner tool.
Nabin Maharjan, Rajendra Banjade, Nobal B. Niraula, Vasile Rus, SemAligner: A Method and Tool for Aligning Chunks with Semantic Relation Types and Semantic Similarity Scores, In LREC 2016, Tenth International Conference on Language Resources and Evaluation , May 23-28, 2016, Portorož, Slovenia.