DISCRIMINATIVE-MODEL LEARNING DEVICE, DISCRIMINATIVE-MODEL LEARNING METHOD, AND DISCRIMINATIVE-MODEL LEARNING PROGRAM

摘要

Provided is a discriminative-model learning device capable of maintaining the fitting of data while efficiently learning a discriminative model which reflects user model knowledge and domain knowledge expressing analysis intent. A query-candidate storage means (81) stores query candidates which are models to which domain knowledge expressing user intent is to be assigned. A regularization-function generation means (82) generates a regularization function, which is a function expressing compatibility with the domain knowledge, on the basis of the domain knowledge assigned to the query candidates. A model-learning means (83) learns a discriminative model by optimizing a function defined using a loss function and a regularization function that are pre-set for each discriminative model.