A statistical confidence measure framework for probabilistic parsing

The thesis introduces the Confidence Measure framework to the parsing world. They are widely used in Automatic Speech Recognition and Statistical Machine Translation, where they are used to asses the confidence in the correctness of each word in a given output sentence. In parsing, confidence refer to each edge (or constituent) of a given tree.Sánchez Sáez, R. (2008). A statistical confidence measure framework for probabilistic parsing. http://hdl.handle.net/10251/13288Archivo delegad