Ígor Assis Braga,  Maria Carolina Monard. 

One of the challenges in machine learning is the automatization of empirical inference processes. In this paper we deal with selective inference, which is an inference type that remains unexplored in machine learning. Des- pite the possibility of using more general inference algorithms for solving se- lective inference, we argue that an algorithm for solving it directly can achieve success even when more general inference algorithms fail. We report on initial experiments conducted on text data sets that support this idea.

http://www.lbd.dcc.ufmg.br/colecoes/enia/2011/0030.pdf

Processando consulta.