Unlock instant, AI-driven research and patent intelligence for your innovation.

ICD code prediction method and system based on joint learning and denoising mechanism

A prediction method and code technology, applied in the field of ICD code prediction method and system based on joint learning and denoising mechanism, can solve problems such as inaccurate matching of ICD codes, difficult classification and coding of key information, lengthy electronic medical records, etc., and achieve improvement The effect of predicting performance, enhancing learning ability, and expanding horizons

Pending Publication Date: 2022-02-08
DALIAN MARITIME UNIVERSITY
View PDF0 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] (3) Electronic medical records are mostly lengthy documents, and it is difficult to extract key information from them for classification and coding
[0008] The problems in the above three aspects will lead to the inaccurate matching between the automatically assigned ICD code and the electronic medical record

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • ICD code prediction method and system based on joint learning and denoising mechanism
  • ICD code prediction method and system based on joint learning and denoising mechanism
  • ICD code prediction method and system based on joint learning and denoising mechanism

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0061] like figure 1 As shown, the present embodiment provides a method for predicting ICD codes based on a joint learning and denoising mechanism, the method comprising:

[0062] S1: Preprocessing the acquired electronic medical record dataset and ICD code description file:

[0063] The acquired electronic medical record dataset is divided into training set, validation set and test set. The ICD codes are divided into 19 categories according to categories.

[0064] Design the preprocessing program of electronic medical records and ICD code description files, convert the original corpus into the input that the deep learning network model can accept, delete the stop words in the electronic medical records, and construct the model-specific thesaurus file of the data set.

[0065] In the specific implementation process, the MIMIC data set shown in Table 1 is used. MIMIC is an intensive care data set released by the Computational Physiology Laboratory of the Massachusetts Institu...

Embodiment 2

[0118] Based on the same inventive concept, this embodiment provides an IDC code prediction system based on a joint learning and denoising mechanism, which includes:

[0119] The data preprocessing module is used to obtain the electronic medical record data set and the ICD code description file, and preprocess the electronic medical record data set and the ICD code description file;

[0120] The model establishment and training module is used to establish the ICD code prediction model based on the deep learning network, and is used to train the ICD code prediction model by using the preprocessed electronic medical record data set and the ICD code description file;

[0121] The model prediction module is used to use the trained ICD code prediction model to predict the ICD code of the electronic medical record to be predicted, and obtain the ICD code that matches the electronic case to be predicted.

[0122] Among them, the data preprocessing module includes:

[0123] The data ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses an ICD code prediction method and system based on joint learning and a denoising mechanism, and relates to the technical field of natural language processing. ClinicalBERT is used for pre-training, a prompt-based fine tuning method is designed, the representation effect of a lengthy sentence is improved, and the pre-training speed is increased. A double-path attention mechanism is used for processing documents of electronic medical records and medical codes, two parts of data are considered at the same time, and the unbalanced classification problem is effectively solved. Different attention matrixes are fed to a joint learning module, two weight coefficients are introduced, the two coefficients are adaptively determined, and an attention matrix specific to ICD is constructed through the two coefficients. A novel denoising loss function is designed, a loss threshold is introduced, sample loss is calculated and sorted, samples exceeding the threshold are cut off, samples exceeding a dynamic threshold in the iteration process are discarded, noisy samples are finally recognized and cleaned, and the training quality of a classifier is improved.

Description

technical field [0001] The present invention relates to the technical field of natural language processing, in particular to an ICD code prediction method and system based on a joint learning and denoising mechanism. Background technique [0002] ICD (International Classification Of Diseases, International Classification of Diseases) is an internationally unified disease classification method formulated by the World Health Organization. The combination of the system and the system expressed by the coding method is a common coding method used in hospitals and various medical systems. A number of predefined ICD codes that can be assigned to patient records such as electronic health records (EHRs). These codes represent diagnostic, medication and procedural information during a patient visit. [0003] Traditionally, clinical diagnostic coding is performed by trained coders. ICD coders translate diseases, pathological causes, symptoms, and signs into standard ICD codes, which...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G16H10/60G06Q10/04G06N3/08G06N3/04G06K9/62G06F40/30
CPCG16H10/60G06F40/30G06Q10/04G06N3/08G06N3/044G06N3/045G06F18/211Y02A90/10
Inventor 张益嘉李兴旺李晓博
Owner DALIAN MARITIME UNIVERSITY
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com