UBC_UOS-TYPED: Regression for typed-similarity
Eneko Agirre, Νικόλαος Αλέτρας, Aitor González-Agirre, Germán Rigau, Mark Stevenson · 2013
We approach the typed-similarity task using a range of heuristics that rely on information from the appropriate metadata fields for each type of similarity. In addition we train a linear regressor for each type of similarity. The results indicate that the linear regression is key for good performance. Our best system was ranked third in the task. 1