Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Classifier training method and computer readable storage medium

A training method and classifier technology, applied in the field of classifier training, can solve the problems of low classification results, narrow adaptability, large detection accuracy error, etc., so as to reduce the time required for training, improve the classification effect, and improve the extraction time. Effect

Pending Publication Date: 2020-03-31
GUILIN UNIV OF ELECTRONIC TECH
View PDF1 Cites 12 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, judging from the pain detection results, the classification results have not been high, and it is difficult to achieve the desired effect
[0004] Again, for the selection of pain features, the most widely used is the time-frequency feature, but the time-frequency feature has high signal requirements and a narrow adaptability, which cannot accurately reflect the accuracy of detection; for facial expressions and pain ratings, they are all subjective. The analysis method has a large error in detection accuracy; for heartbeat interval, photoelectric pulse wave amplitude change, surgical blood oxygen volume index, autonomic nervous system state index, etc., the actual operation workload is huge, and it is difficult to achieve cheap, portable and real-time detection

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
  • Classifier training method and computer readable storage medium
  • Classifier training method and computer readable storage medium
  • Classifier training method and computer readable storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0061] In the embodiment proposed by the present invention, such as figure 1 As shown, the training methods of the classifier include:

[0062] Step S102, decomposing the time series of EEG signals containing cold pain information to obtain decomposed subband components;

[0063] Step S104, determining the ratio of the energy value of any subband component to the sum of the energy values ​​of all subband components, the fine compound multiscale dispersion entropy of any subband component, the fine compound multiscale fuzzy entropy and the autoregressive model coefficient;

[0064] Step S106, determining the feature set according to the ratio of the energy value of any subband component to the sum of the energy values ​​of all subband components, fine composite multiscale dispersion entropy, fine composite multiscale fuzzy entropy, and autoregressive model coefficients;

[0065] Step S108, train the classifier according to the feature set to obtain the target classifier.

[0...

Embodiment 2

[0143] In the above embodiment, further comprising: performing dimensionality reduction on the feature set.

[0144] In this embodiment, by reducing the dimensionality of the feature set, the problem of large amount of calculation, calculation time, and excessive resource consumption caused by too many feature dimensions is eliminated, and the impact of too many feature dimensions on feature detection and feature detection is reduced. Classification effects.

[0145] Such as Figure 6 As shown, use a feature dimensionality reduction algorithm for dimensionality reduction, such as PCA (Principal components analysis, principal component analysis), to remove redundant features, and its specific steps include:

[0146] Step 1, get n-dimensional sample set D={x (1) , x (2) ,...x (m)};

[0147] Step 2, centralize all samples: Form a new data set X={s (1) ,s (2) ,...,s (m)};

[0148] Step 3, calculate the covariance matrix XX of the sample T ;

[0149] Step 4, for matrix...

Embodiment 3

[0154] In any of the above embodiments, it also includes: determining at least one of the sensitivity, specificity, positive predictive value and accuracy of the classifier; and according to at least one of the sensitivity, specificity, positive predictive value and accuracy A selection target classifier.

[0155] In this embodiment, the classifier is evaluated by introducing one or more evaluation indicators in sensitivity, specificity, positive predictive value and accuracy, and the target classifier is selected according to the evaluation results, so as to obtain the optimal classifier , so as to ensure the classification accuracy of the classifier.

[0156] Among them, the calculation formulas of sensitivity, specificity, positive predictive value and accuracy are as follows:

[0157]

[0158]

[0159]

[0160]

[0161] Among them, TP means true positive; TN means true negative; FP means false positive; FN means false negative.

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 provides a classifier training method and a computer readable storage medium, and the classifier training method comprises the steps: carrying out the decomposition of a time sequence ofan electroencephalogram signal containing crymodynia information, so as to obtain decomposed sub-band components; determining a ratio of the energy value of any sub-band component to the sum of the energy values of all the sub-band components, a fine composite multi-scale dispersion entropy of any sub-band component, a fine composite multi-scale fuzzy entropy and an autoregression model coefficient; determining a feature set according to the ratio of the energy value of any sub-band component to the sum of the energy values of all the sub-band components, the fine composite multi-scale dispersion entropy, the fine composite multi-scale fuzzy entropy and the autoregression model coefficient; and training the classifier according to the feature set to obtain the target classifier. The classification effect of the classifier obtained through training according to the feature variables on the crymodynia information is better, and the reliability of the diagnosis result of the neurologicaldiseases is improved.

Description

technical field [0001] The present invention relates to the field of classifier training, in particular to a classifier training method and a computer-readable storage medium. Background technique [0002] In related technical proposals, acute pain has been widely studied as a way to induce brain activity signals, and as a laboratory method for clinical diagnosis of pain-induced diseases. At present, most of the pain detection is based on the collection of scalp EEG, (electroencephalograph, electroencephalogram), fMRI, (functional magnetic resonance imaging, functional magnetic resonance imaging), EMG, namely electromyography, and the application of electronic instruments to record muscle rest Or the electrical activity during contraction, and the method of applying electrical stimulation to examine the nerve, muscle excitation and conduction function and ECG, that is, electrocardiogram, electrocardiogram and other physiological signals for simple physiological feature extra...

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): G06K9/62
CPCG06F18/2135G06F18/24G06F18/214
Inventor 杨道国耿道双蔡苗张国旗郝卫东
Owner GUILIN UNIV OF ELECTRONIC TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com