NEWLY-ARRIVED INFORMATION SEARCH SYSTEM, SEARCH METHOD, AND PROGRAM

WIPO Patent Application WO/2018/021227

A1

A result obtained by an SDI search tends to depend on a user's proficiency, and may include an unnecessary noise component, or may possibly fail to extract necessary publications. This search system includes: a model constructing unit which constructs a learning model on the basis of teacher data which represents a relationship between input data about at least one publication and output data about at least one publication corresponding to the input data; a newly-arrived information input unit through which newly-arrived information about publications published in a prescribed period is input, as the input data, to the learning model; an output result providing unit which provides, to a user terminal, a result output from the learning model according to the input of the newly-arrived information; a receiving unit which receives, from the user terminal, teacher information which is a source of the teacher data; and a teacher data input unit which inputs, to the model constructing unit, the teacher data according to the teacher information.

WANG LIANG (JP)
KATO REO (JP)
NAGANUMA SATOSHI (JP)
TAJIMA RYUTARO (JP)
HIROMOTO HIDEHISA (JP)
TAKANO SEIJI (JP)

PCT/JP2017/026648

February 01, 2018

July 24, 2017

Click for automatic bibliography generation  

NOMURA RES INST LTD (JP)

G06F17/30; G06N99/00

JPH09101969A	1997-04-15
JP2001344276A	2001-12-14
JP2004118877A	2004-04-15
JP2004005751A	2004-01-08

A MATI, GIANNI ET AL.: "Probabilistic learning for selective dissemination of information", INFORMATION PROCESSING & MANAGEMENT, vol. 35, no. 5, September 1999 (1999-09-01), pages 633 - 654, XP004178704

INABA, Yoshiyuki et al. (JP)

View/Download PDF      PDF Help

Previous Patent: PREFORM PRODUCTION APPARATUS AND PREFORM PRODUCTION METHOD

Next Patent: ELECTRICAL CONNECTION COMPONENT