Data prediction method and electronic device

By jointly modeling doctor behavior and patient state, and using an attention weight matrix to achieve information interaction, a joint output integrating behavioral intention and state evolution information is generated. This solves the problem of disconnect between doctor behavior and patient state in existing technologies and improves the accuracy and consistency of data prediction.

CN122392782APending Publication Date: 2026-07-14BEIJING UNIV OF POSTS & TELECOMM
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
BEIJING UNIV OF POSTS & TELECOMM
Filing Date
2026-03-05
Publication Date
2026-07-14

AI Technical Summary

Technical Problem

Existing data prediction technologies in clinical practice ignore the bidirectional causal relationship and mutual influence between physician behavior and patient physiological state, resulting in a serious disconnect between the generated auxiliary suggestions and the patient's latest clinical status, lacking foresight and contextual synergy.

Method used

By acquiring doctor behavior data sequences and patient clinical data sequences as dual-channel inputs, a joint modeling foundation is established. Information interaction is achieved using an attention weight matrix, and a joint output integrating behavioral intention and state evolution information is generated through a fusion prediction model.

Benefits of technology

It significantly improves the accuracy of data prediction, ensures the synergy and consistency of prediction, and overcomes the shortcomings of existing methods that lack dynamic correlation between behavioral data prediction and clinical data prediction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122392782A_ABST
    Figure CN122392782A_ABST
Patent Text Reader

Abstract

The application relates to the technical field of data processing, in particular to a data prediction method and an electronic device, wherein the method comprises designing a prediction model, the prediction model simultaneously receives a behavior data sequence of a first entity and a clinical data sequence of a second entity, respectively outputs initial predicted behavior data and initial predicted clinical data, and realizes bidirectional information interaction through an attention mechanism: an attention weight is calculated based on an association relationship of a hidden state vector, and a context vector is generated; the hidden state vector and the context vector are fused to form joint features, and the joint features are used to correct the initial prediction of each channel, and finally the correction results of the two channels are fused to obtain joint prediction. Through the introduction of the cross-channel attention interaction and joint correction mechanism, the accuracy of data prediction is improved.
Need to check novelty before this filing date? Find Prior Art