... with training a model using a base-line set of features, and then add each remainingfeature individually, recording the gain (w.r.t. the f-measure (f (0.5), to be precise)), and choosing thebest-performing ... relations (as in (1-b) and (1-c)above). Structural information (iqu) further im-proves the model; however, the remaining featuresonly seem to add interfering information, for perfor-mance using the ... Proceedings of the 43rd Annual Meeting of the ACL, pages 247–254,Ann Arbor, June 2005.c2005 Association for Computational LinguisticsTowards Finding and Fixing Fragments: Using ML to...