STANDARD MODEL GENERATION APPARATUS AND STANDARD MODEL GENERATION METHOD

摘要

PROBLEM TO BE SOLVED: To provide a highly accurate standard model used for pattern recognition such as speech recognition, character recognition, and image recognition based upon probabilistic models such as a hidden Markov model, Bayes theory and linear discrimination analysis, intention understanding by a probabilistic model of a Bayesian net, and data mining by a probabilistic model.;SOLUTION: A standard model generation apparatus is provided with a reference model preparation part 102 which prepares one or more reference models, a reference model storage part 103 which stores the reference models 121 prepared by the reference model preparation part 102, and a standard model generation part 104 which generates a standard model 122 by calculating statistics of the standard model so that the probability or likelihood of one or more reference models stored in the reference model storage part 103 is maximized or relatively maximized.;COPYRIGHT: (C)2005,JPO&NCIPI