model learning device, related information extraction device, related information prediction apparatus, model learning method, related information extraction method, related information prediction method and program

摘要

PROBLEM TO BE SOLVED: To provide a model learning device capable of acquiring related data from each of a plurality of data sets with different dimensions (the number of data items), and to provide a related information extraction device, a related information prediction device, a method for learning models, a method for extracting related information, a method for predicting related information, and a program.SOLUTION: A learning unit 20 estimates three types of parameters, a probability set that is a set of probabilities corresponding to data in each data set, a latent vector set, and a transformation matrix set that is a set of multiple transformation matrices corresponding to each of each data set such that a model is optimized to the plurality of input data sets. The model is defined that data in each data set in the plurality of data sets are generated at a predetermined probability depending on one latent vector in a predetermined latent vector set and a transformation matrix corresponding thereto.