Title:
MODEL VARIABLE CANDIDATE GENERATION DEVICE AND METHOD
Document Type and Number:
WIPO Patent Application WO/2018/186090
Kind Code:
A1
Abstract:
The present invention makes it possible to quickly narrow down the variables of a model in data analysis. A model variable candidate generation device for generating an explanatory variable candidate that is made a candidate for an explanatory variable in generating a prediction model, the device having a data input unit for inputting data for analysis that includes one or more items for each entry and has item values for the items, a first item determination unit for preliminarily setting the natures of the items included in the data for analysis as first item natures, a data nature determination unit for determining data natures that are the natures of the data for analysis on the basis of the first item natures of the items included in the data for analysis, a second item determination unit for determining the natures of the items included in the data for analysis as second item natures on the basis of the data natures of the data for analysis, and a variable candidate generation unit for generating an explanatory variable candidate on the basis of the second item natures of the items included in the data for analysis by selecting from the items or processing the items.
More Like This:
Inventors:
FUJIMOTO KOJI (JP)
SHIBAHARA KAZUTOMO (JP)
KOREKAWA TAKASHI (JP)
SHIBAHARA KAZUTOMO (JP)
KOREKAWA TAKASHI (JP)
Application Number:
PCT/JP2018/008465
Publication Date:
October 11, 2018
Filing Date:
March 06, 2018
Export Citation:
Assignee:
TENSOR CONSULTING CO LTD (JP)
International Classes:
G06F17/30; G06Q10/04
Domestic Patent References:
| WO2010082322A1 | 2010-07-22 |
Foreign References:
| JP2002358411A | 2002-12-13 | |||
| JP2007329415A | 2007-12-20 | |||
| JP2015224389A | 2015-12-14 | |||
| JP2013065084A | 2013-04-11 | |||
| JP2013152656A | 2013-08-08 | |||
| JP2008158748A | 2008-07-10 | |||
| JP6069460B1 | 2017-02-01 | |||
| JP2016004525A | 2016-01-12 |
Other References:
TANIOKA, HIDEO ET AL.: "AI technologies in financial business industry", JOURNAL OF JAPANESE SOCIETY FOR ARTIFICIAL INTELLIGENCE, vol. 17, no. 2, 1 March 2002 (2002-03-01), pages 214 - 221, XP9516329
FUJINO, HIDENORI: "Analysis of causal relationship (1): regression analysis & multiple regression analysis", JOURNAL OF HUMAN INTERFACE SOCIETY: HUMAN INTERFACE, vol. 15, no. 2, 25 May 2013 (2013-05-25), pages 141 - 149, XP009517714
UEDA, TAICHIRO: "A function on variable selection in discriminant analysis and its applications", BULLETIN OF THE COMPUTATIONAL STATISTICS OF JAPAN, vol. 8, no. 2, 31 May 1996 (1996-05-31), pages 171 - 175, XP009516883, DOI: 10.20551/jscswabun.8.2_171
FUJIMAKI, RYOHEI ET AL.: "Challenge for automation and standardization of analysis process", JOURNAL OF DIGITAL PRACTICES, vol. 6, no. 3, November 2016 (2016-11-01), pages 198 - 206, XP009516884, ISSN: 2188-4390
See also references of EP 3608802A4
FUJINO, HIDENORI: "Analysis of causal relationship (1): regression analysis & multiple regression analysis", JOURNAL OF HUMAN INTERFACE SOCIETY: HUMAN INTERFACE, vol. 15, no. 2, 25 May 2013 (2013-05-25), pages 141 - 149, XP009517714
UEDA, TAICHIRO: "A function on variable selection in discriminant analysis and its applications", BULLETIN OF THE COMPUTATIONAL STATISTICS OF JAPAN, vol. 8, no. 2, 31 May 1996 (1996-05-31), pages 171 - 175, XP009516883, DOI: 10.20551/jscswabun.8.2_171
FUJIMAKI, RYOHEI ET AL.: "Challenge for automation and standardization of analysis process", JOURNAL OF DIGITAL PRACTICES, vol. 6, no. 3, November 2016 (2016-11-01), pages 198 - 206, XP009516884, ISSN: 2188-4390
See also references of EP 3608802A4
Attorney, Agent or Firm:
WILLFORT INTERNATIONAL PATENT FIRM (JP)
Download PDF: