GM-RKB WTEC System's False Positive (FP) Classification
Jump to navigation
Jump to search
A GM-RKB WTEC System's False Positive (FP) Classification is a False Positive Classification in which a non-wikitext error is wrongly detected and corrected as an error by the GM-RKB WTEC System.
- Context:
- It usually leads to the introduction a new wikitext error by the system in the attempt of correcting a wrong error detection.
- …
- Example(s):
- …
- Counter-Example(s):
- See: GM-RKB WikiText Error Correction (WTEC) Task, GM-RKB WTEC System's Performance Evaluation Metric, WikiText Error Correction (WTEC) System.
References
2020
- (Melli et al., 2020) ⇒ Gabor Melli, Abdelrhman Eldallal, Bassim Lazem, and Olga Moreira. (2020). “GM-RKB WikiText Error Correction Task and Baselines.”. In: Proceedings of LREC 2020 (LREC-2020).
- QUOTE: Each character in the WEC system's output text is grouped into one of the following five outcomes:
- True Positive (TP): the source character was an error; the model correctly fixed it.
- False Positive (FP): although the source character was not an error; however, the model tried to fix it introducing a new error instead.
- True Negative (TN): the source character was not an error; the model correctly disregarded it as an error and did not try to fix it.
- False Negative (FN): the source character was an error; however, the model failed to detect it and processed the error as the correct character.
- Detected Not-Fixed (DN): the source character was an error; the model successfully detected the error but failed to correct it, or used the wrong character to fixed. (e.g. the correct repair was changing 'a' to 'b' but the model changed 'b' to 'c' instead of 'a').
- QUOTE: Each character in the WEC system's output text is grouped into one of the following five outcomes: