Title:
DEVICE AND METHOD FOR INFERRING DEPRESSIVE STATE AND PROGRAM FOR SAME
Document Type and Number:
WIPO Patent Application WO/2020/122227
Kind Code:
A1
Abstract:
Provided is a novel method for inferring a depressive state using a wearable device. Each data type of biological data for several days from a plurality of subjects is converted into data for each predetermined unit time such as one hour (S201). Next, the quantile of the distribution of obtained sample data is determined for each subject and each data type (S202). The standard deviation of the distribution of the obtained sample data is also calculated for each subject and each data type (S203). The Pearson correlation coefficient is calculated for each combination of data types for each subject (S204). Next, an inference model for the problem of determining whether a subject is classified as being in a depressive state or not is trained by machine leaning using teacher data in which the feature amounts of the quantile, the standard deviation, and the Pearson correlation coefficient extracted from the biological data of each subject are used as input vectors and evaluation of the depressive state for each subject by an expert such as a doctor is used as a label (S206).
Inventors:
KISHIMOTO TAISHIRO (JP)
TAZAWA YUKI (JP)
KUO-CHING LIANG (JP)
FUJITA TAKANORI (JP)
YOSHIMURA MICHITAKA (JP)
KITAZAWA MOMOKO (JP)
MIMURA MASARU (JP)
TAZAWA YUKI (JP)
KUO-CHING LIANG (JP)
FUJITA TAKANORI (JP)
YOSHIMURA MICHITAKA (JP)
KITAZAWA MOMOKO (JP)
MIMURA MASARU (JP)
Application Number:
PCT/JP2019/048904
Publication Date:
June 18, 2020
Filing Date:
December 13, 2019
Export Citation:
Assignee:
UNIV KEIO (JP)
International Classes:
A61B10/00; A61B5/01; A61B5/02; A61B5/16; A61B5/22; G16H50/20
Foreign References:
| US20170238858A1 | 2017-08-24 | |||
| JP2018524137A | 2018-08-30 | |||
| JP2018015327A | 2018-02-01 | |||
| JP2016163698A | 2016-09-08 | |||
| CN108888281A | 2018-11-27 | |||
| KR101911516B1 | 2018-10-24 |
Attorney, Agent or Firm:
OTANI Kan (JP)
Download PDF: