Japanese Case Analysis Based on Machine Learning Method that Uses Borrowed Supervised Data
Masaki Murata, Hitoshi Isahara · 2006
We developed a new machine learning method, in which supervised data are borrowed from corpora that do not have annotated tags related to the problems to be solved. We also developed a second machine learning method that uses both borrowed supervised data and normal supervised data. Both methods can be used for any type of ellipsis resolution. We demonstrate the effectiveness of these methods for Japanese case analysis.