DISEASE PREDICTION DEVICE, PREDICTION MODEL GENERATION DEVICE, AND DISEASE PREDICTION PROGRAM

Title:

DISEASE PREDICTION DEVICE, PREDICTION MODEL GENERATION DEVICE, AND DISEASE PREDICTION PROGRAM

Document Type and Number:

WIPO Patent Application WO/2021/106825

Kind Code:

A1

Abstract:

This device, for extracting acoustic feature values from data in conversation audio data, performing machine learning, and predicting the disease level of a subject on the basis of a disease prediction model generated on the basis of said machine learning, is provided with a matrix calculation unit 23 which calculates a spatial delay matrix that uses relation values between acoustic feature values of different types, and a matrix analysis unit 24 which calculates a matrix analysis value from the spatial delay matrix; by calculating the DCCA coefficient and/or mutual information amount as the relation value between the acoustic feature values of different types, a relation value is obtained that reflects the non-linear and non-stationary relation between feature values, and this enables predicting the disease level of a subject on the basis of said relation value.

Inventors:

KISHIMOTO TAISHIRO (JP)
LIANG KUO-CHING (JP)
YOSHIMURA MICHITAKA (JP)
YOSHIMURA MOMOKO (JP)
FUJITA TAKANORI (JP)
MIMURA MASARU (JP)

Application Number:

PCT/JP2020/043563

Publication Date:

June 03, 2021

Filing Date:

November 24, 2020

Export Citation:

Click for automatic bibliography generation Help

Assignee:

UNIV KEIO (JP)

International Classes:

G16H50/20; A61B10/00; G06F17/18; G10L15/10; G10L25/66

Domestic Patent References:

WO2002087434A1

2002-11-07

Foreign References:

JP2004240394A	2004-08-26
JP2002306492A	2002-10-22
JP2017532082A	2017-11-02

Other References: