A data-length adaptive method for electrocardiogram classification

A technology of data length and classification method, applied in neural learning methods, instruments, biological neural network models, etc., can solve problems such as low accuracy, few data sets, and uneven data set quality, and achieve easy training and accurate classification , the effect of rapid convergence

Active Publication Date: 2021-10-08
SICHUAN UNIV
View PDF10 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to solve the problem of low accuracy of the existing electrocardiogram training results due to the lack of electrocardiogram data sets and the uneven quality of the data sets in the prior art. This application provides a data length self-adaptive electrocardiogram classification and training method

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
  • A data-length adaptive method for electrocardiogram classification
  • A data-length adaptive method for electrocardiogram classification
  • A data-length adaptive method for electrocardiogram classification

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0107] On the basis of the above, the present embodiment of the above detailed analysis:

[0108] 1. The raw data is processed to obtain a second section 24 by a dynamic segmentation and connected for subsequent training algorithm

[0109] (1-1) and connected dynamic segmentation

[0110] Unbalanced raw ECG data, the length of each ECG inconsistent record. In order to solve these problems, a method of dynamic segmentation and connected as a sampling method. The specific method is as follows:

[0111] For ECG Records longer than 24 seconds, 24 seconds, as a new random data is cut. For shorter than 24 seconds of ECG recording, first randomly cut into three segments, a length of 8 seconds. Specification of each segment according to the formula:

[0112]

[0113] Segment is regarded as a vector X = (x 1 , X 2 , ..., x t ) Sequences, and these three sections are spliced ​​together in accordance with the peak amplitude of the R (Rpeakamplitude). Where R is the peak amplitude of the EC...

Embodiment 2

[0145] On the basis of the first example, the above results are evaluated, and the F score is used to measure the accuracy of the classification problem category, for normal rhythm, AF rhythm, other rhythm, and noise:

[0146]

[0147]

[0148]

[0149]

[0150] The last score is as follows:

[0151]

[0152] Among them, NN - is actually normal data, and the model forecast results are also normal samples;

[0153] ΣN - Total number of samples of normal data;

[0154] The number of samples that are predicted as normal data is predicted;

[0155] AA- is actually AF rhythm data, and the model forecast results are also the number of samples of the AF rhythm;

[0156] The total number of samples of σa - AF rhythm data;

[0157] Σa - The total number of samples predicted as AF rhythm data is predicted;

[0158] OO - actually other rhythm data, model prediction results are also the number of samples of other rhythms;

[0159] ΣO - The total number of samples of other rhythm dat...

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 a data length self-adaptive electrocardiogram classification method, which aims to solve the problem of low accuracy of the existing electrocardiogram training results caused by the lack of electrocardiogram data sets and the uneven quality of the data sets in the prior art; the application is adopted The original data is dynamically segmented and connected to obtain the standard time-length ECG segment data set. This application dynamically segments and connects the obtained ECG data set to generate a large number of additional training data sets, and collects normal ECG data, Atrial fibrillation data, other data, and noise data are four types of data sets, which maintain the balance of the data set, thereby improving the robustness and generalization ability of the later training, and thus making the classification of the input set more accurate; this application applies In the related field of ECG classification training.

Description

Technical field [0001] The present invention relates to an electrocardiogram classification training related art, and more particularly to an electrocardiogram classification and training method of data length adaptive. Background technique [0002] With the development of social economy, the changes in residents' lifestyle, the process of population aging, etc., the dangers of cardiovascular disease continue to rise, the prevalence of cardiovascular disease in China, the risk of cardiovascular disease, the risk of rising, and the death rate is high About 3.5 million people have died of cardiovascular diseases every year. Overall, the prevalence and mortality rate of cardiovascular disease in my country is still in the rise. Cardiovascular disease occupies more than 40% of the death of residents, for the first death cause of my country's residents. The arrhythmia is a kind of cardiovascular disease, and its treatment and prevention have become more concerned about people. [0003...

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 Patents(China)
IPC IPC(8): G06K9/62G06N3/04G06N3/08
CPCG06N3/082G06N3/045G06F18/211G06F18/2415G06F18/214
Inventor 吕建成陈尧刘大一恒李茂
Owner SICHUAN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap