摘要 |
PROBLEM TO BE SOLVED: To suppress increase of processing load even in improving learning accuracy of so-called "supervised learning". SOLUTION: The data processor comprises a first unsupervised learning processing means 13, which classifies data by unsupervised learning to a first data group 21 to obtain a first classification data group 23 in which the data order number of the first data group 21 is reduced; a second unsupervised learning processing means 14 which classifies data by unsupervised learning to a second data group 22 to obtain a second classification data group 24 in which the data order number of the second data group 22 is reduced; and a supervised learning processing means 15 which specifies a mapping relation between the first data group 23 obtained by the first means 13 and the second data group 24 obtained by the second means 14 by supervised learning using the first data group 23 and the second data group 24 as a teacher. COPYRIGHT: (C)2008,JPO&INPIT
|