Comma checking in Danish

This paper describes research in using the Brill tagger (Brill 94,95) to learn to identify incorrect commas in Danish. Trained on a part-of-speech tagged corpus of 600,000 words, the system identifies incorrect commas with a precision of 91% and a recall of 77%. The system was developed by randomly inserting commas in a text, which were tagged as incorrect, while the original commas were tagged as correct. Then the tagger was trained to recognize the contexts in which incorrect commas occur. In what follows, we first describe the corpora and tag sets used in this research, and give background on the Brill Tagger. We then describe the methodology for learning to identify comma errors, and then we examine some of the principles that the system learned to identify comma errors. Finally, test results are presented, and we discuss plans for future research. The method used here is quite general, and could be applied fairly directly to a wide range of grammar checking problems, in Danish or other languages.