INFORMATION PROCESSING DEVICE, METHOD, PROGRAM, IC CHIP, AND LEARNED MODEL

WIPO Patent Application WO/2018/168971

A1

[Problem] To achieve highly accurate prediction processing, even when data to be learned includes noise or the like. [Solution] An information processing device that generates prediction output that corresponds to input data. The information processing device comprises: an input-node specifying unit that, on the basis of the input data, specifies, at each layer of a learning tree structure from the beginning to the end of the learning tree structure, an input node that corresponds to the input data; a reliability-index acquisition unit that acquires a reliability index that is obtained by means of prescribed learning processing and indicates prediction accuracy; an output-node specifying unit that, on the basis of the reliability index obtained by the reliability-index acquisition unit, specifies, from the input nodes that correspond to the input data, an output node that will be the basis for the generation of the prediction output; and a prediction-output generation unit that generates the prediction output on the basis of data to be learned that is included in a state space that corresponds to the output node specified by the output-node specifying unit.

KIM CHYON HAE (JP)
NUMAKURA AKIO (JP)
SUGAWARA YASUHIRO (JP)
IDESAWA JUNICHI (JP)

PCT/JP2018/010089

September 20, 2018

March 14, 2018

Click for automatic bibliography generation  

NATIONAL UNIV CORPORATION IWATE UNIV (JP)
AISING LTD (JP)

G06N99/00; G06F17/30

US20050144147A1	2005-06-30
JP2016173686A	2016-09-29

KIM C ET AL : "Error based prediction algorithm in dynamics learning tree", IPSJ SIG TECHNICAL REPORT, vol. 2017-MPS-12, no. 25, 20 February 2017 (2017-02-20), pages 1 - 5, XP009515865, ISSN: 2188-8833

IIZUKA Shin-ichi et al. (JP)

View/Download PDF      PDF Help

Previous Patent: HYDRAULIC CONTROL DEVICE

Next Patent: SURFACE TREATMENT AGENT COMPRISING SILANE COMPOUND CONTAINING PERFLUORO(POLY)ETHER GROUP, AND PELLET...