%0 Conference Paper %1 derczynski-2016-complementarity %A Derczynski, Leon %B Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC'16) %C Portoroz, Slovenia %D 2016 %I European Language Resources Association (ELRA) %K Complementarity instituteclustering %P 261--266 %T Complementarity, F-score, and NLP Evaluation %U https://www.aclweb.org/anthology/L16-1040 %X This paper addresses the problem of quantifying the differences between entity extraction systems, where in general only a small proportion a document should be selected. Comparing overall accuracy is not very useful in these cases, as small differences in accuracy may correspond to huge differences in selections over the target minority class. Conventionally, one may use per-token complementarity to describe these differences, but it is not very useful when the set is heavily skewed. In such situations, which are common in information retrieval and entity recognition, metrics like precision and recall are typically used to describe performance. However, precision and recall fail to describe the differences between sets of objects selected by different decision strategies, instead just describing the proportional amount of correct and incorrect objects selected. This paper presents a method for measuring complementarity for precision, recall and F-score, quantifying the difference between entity extraction approaches.

@inproceedings{derczynski-2016-complementarity, abstract = {This paper addresses the problem of quantifying the differences between entity extraction systems, where in general only a small proportion a document should be selected. Comparing overall accuracy is not very useful in these cases, as small differences in accuracy may correspond to huge differences in selections over the target minority class. Conventionally, one may use per-token complementarity to describe these differences, but it is not very useful when the set is heavily skewed. In such situations, which are common in information retrieval and entity recognition, metrics like precision and recall are typically used to describe performance. However, precision and recall fail to describe the differences between sets of objects selected by different decision strategies, instead just describing the proportional amount of correct and incorrect objects selected. This paper presents a method for measuring complementarity for precision, recall and F-score, quantifying the difference between entity extraction approaches.}, added-at = {2021-01-19T15:14:03.000+0100}, address = {Portoro{\v{z}}, Slovenia}, author = {Derczynski, Leon}, biburl = {https://www.bibsonomy.org/bibtex/2a33cbe69340ab6b3d8730dfa882f8077/parismic}, booktitle = {Proceedings of the Tenth International Conference on Language Resources and Evaluation ({LREC}'16)}, description = {Complementarity, F-score, and NLP Evaluation - ACL Anthology}, interhash = {0d30a61c7e3ee1835b39837655464ad6}, intrahash = {a33cbe69340ab6b3d8730dfa882f8077}, keywords = {Complementarity instituteclustering}, month = may, pages = {261--266}, publisher = {European Language Resources Association (ELRA)}, timestamp = {2021-01-19T15:14:03.000+0100}, title = {Complementarity, {F}-score, and {NLP} Evaluation}, url = {https://www.aclweb.org/anthology/L16-1040}, year = 2016 }