SELECTION OF VOICE FEATURE FOR ESTABLISHING MODEL FOR USE IN DETECTION OF SYMPTOM

Title:

SELECTION OF VOICE FEATURE FOR ESTABLISHING MODEL FOR USE IN DETECTION OF SYMPTOM

Document Type and Number:

Japanese Patent JP2023052224

Kind Code:

A

Abstract:

To provide a method of establishing a mathematical model for use in detection of a symptom to improve performance of the model.SOLUTION: A method according to the present invention has a step 510 of acquiring a training corpus including a voice data item, a step 220 of acquiring voice recognition result with respect to each of the voice data items, a step 210 of calculating a plurality of acoustic features with respect to each of the voice data items in the training corpus, a step 230 of calculating a plurality of language features with respect to each of the voice data items in the training corpus by processing the voice recognition result, a step 520 of calculating a feature selection score with respect to each of features of the plurality of acoustic features and to each of features of the plurality of language features, a step 540 of selecting a plurality of features from the plurality of acoustic features and the plurality of language features by using the feature selection score, and a step 550 of training a mathematic model for use in detection of a symptom by using the plurality of features selected with respect to each of the voice data items in the training corpus.SELECTED DRAWING: Figure 5

More Like This:

JP7052866	Self-training data sorting device, estimation model learning device, self-training data sorting method, estimation model learning method, and program
WO/2023/227208	ONLINE RETRAINING OF PHYSICAL LAYER TRANSCEIVERS BASED ON MACHINE LEARNING
WO/2021/117921	METHOD FOR EVALUATING INFERENCE CLASSIFICATION PERFORMANCE OF EACH LAYER OF ARTIFICIAL INTELLIGENCE MODEL INCLUDING MULTIPLE LAYERS, AND EVALUATION DEVICE THEREFOR

Inventors:

Kim, jeongwon
Kwon, Nami
Henry O'Connell
Philip, Walstad
Kevin Shenbin, Yang

Application Number:

JP2023000544A

Publication Date:

April 11, 2023

Filing Date:

January 05, 2023

Export Citation:

Click for automatic bibliography generation Help

Assignee:

Canary Speech, LRC

International Classes:

G06N20/00; A61B5/00; G10L15/00; G10L25/30; G10L25/66

Attorney, Agent or Firm:

Osamu Yamamoto
Toru Miyamae
Junichi Matsuo

Previous Patent: Systems and methods for tissue removal

Next Patent: Assembly process for automated footwear platform