Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
MODULAR TIME SERIES DATA PREDICTION DEVICE, MODULAR TIME SERIES DATA PREDICTION METHOD, AND PROGRAM
Document Type and Number:
WIPO Patent Application WO/2022/239609
Kind Code:
A1
Abstract:
Provided is a modular time series data prediction device that logically explains a prediction result and that enables easy adjustment of the prediction result. The device according to an embodiment includes: a function 2 for collecting and storing, at a prescribed cycle, time series data for a plurality of process variables, selecting an output variable to be predicted from among the plurality of process variables, and selecting an input variable candidate from among the plurality of process variables; a function 4 for defining a 1-input/1-output pairwise prediction model by using data for identifying a plurality of input variables and an output variable extracted from the time series data; a function 5 for defining a method for synthesizing prediction values of a plurality of pairwise prediction models; a function 6 for inputting prediction data for the plurality of input variables into a plurality of pairwise prediction models, and calculating prediction values corresponding respectively to the plurality of input variables; and a function 7 for calculating a prediction value for the output variable by synthesizing the plurality of prediction values according to the defined synthesis method.

Inventors:
YAMANAKA OSAMU (JP)
HIRAOKA YUKIO (JP)
Application Number:
PCT/JP2022/018118
Publication Date:
November 17, 2022
Filing Date:
April 19, 2022
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
TOSHIBA KK (JP)
TOSHIBA INFRASTRUCTURE SYSTEMS & SOLUTIONS CORP (JP)
International Classes:
G05B23/02
Foreign References:
JP2018116545A2018-07-26
JP2016507113A2016-03-07
JP2012515902A2012-07-12
JP2005222444A2005-08-18
JP2004062440A2004-02-26
JP6261960B22018-01-17
Attorney, Agent or Firm:
KURATA, Masatoshi et al. (JP)
Download PDF: