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

Brain network clustering method based on local attributes and topological structure

A technology of local attributes and topological structures, which is applied to computer components, character and pattern recognition, instruments, etc., and can solve problems such as inaccurate labeling

Active Publication Date: 2019-03-01
TAIYUAN UNIV OF TECH
View PDF2 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a brain network clustering method based on local attributes and topological structures, to overcome the problem of inaccurate labeling in supervised learning, and to make the brain network similarity clustering results more accurate

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
  • Brain network clustering method based on local attributes and topological structure
  • Brain network clustering method based on local attributes and topological structure
  • Brain network clustering method based on local attributes and topological structure

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0051] ⑴Experimental data

[0052] Taking Alzheimer's disease and normal elderly people as examples, cluster analysis was performed on the brain networks of the two types of subjects; the research data came from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, including 28 Alzheimer's patients (Age: 72.2±7.5, female: 17, Mini-Mental State Examination (MMSE): 22.8±2.5, Clinical Dementia Rating Scale (CRD): 0.84±0.23), 28 normal elderly people (age: 74.3±6.3 , female: 17, MMSE: 28.9±1.3, CRD: 0±0);

[0053] ⑵Experimental process

[0054] according to figure 1 In the process shown, the following steps are used to cluster the brain networks of 56 subjects:

[0055] Step S1: Preprocessing the fMRI images of the brain, then segmenting the brain regions, and extracting the average time series of each brain region;

[0056] Dparsf software was used to preprocess the functional magnetic resonance images of the brain. The preprocessing steps specifically included: re...

Embodiment 2

[0093] The weight value of step 3 in embodiment 1 is taken as 0.7, and each similarity obtained for step 3: S Atr (G,H),S Str (G,H) and S(G,H) are clustered using a clustering algorithm, and the precision rate, recall rate and F1 value of the clustering results are calculated and tested;

[0094] The precision rate, recall rate and F1 value of each clustering result are as follows: figure 2 shown by figure 2 It can be seen that the weighted combination of local attribute similarity and topology similarity has the best clustering effect of brain network similarity, with a precision rate of 0.64, a recall rate of 0.62, and an F1 of 0.63.

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 brain network clustering method based on local attribute and topological structure. The method is carried out according to the following steps: firstly, the brain function network is pretreated and the average time series of each brain region is extracted; then the Pearson correlation coefficients are calculated to construct the unbiased brain function network. The similarity of brain function network is calculated. At last, the similarity of brain functional network is clustered and the clustering results are tested. The invention obtains a higher precision similarity by weighted fusion of local attribute similarity degree and topological structure similarity degree of the brain function network. After clustering the similarity degree based on the local attributeand topological structure, the clustering result obtained is accurate and without bias.

Description

technical field [0001] The invention belongs to the technical field of machine learning, in particular to a brain network clustering method based on local attributes and topological structures. Background technique [0002] At present, machine learning, as an important tool for brain network analysis, has become a new research hotspot in the field of brain network analysis in recent years because it can learn rules from data and predict unknown data. Supervised learning (classification) and unsupervised learning (clustering). [0003] Most of the current research uses supervised learning, that is, using labeled training data to train the classification model, and then using the classification model to classify the test data; however, since the labeling of the data is performed by professionals based on some prior knowledge Labeling is subjective, and errors may occur during the labeling process, which will eventually affect the classification results and reduce the accuracy...

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
IPC IPC(8): G06K9/62G06K9/34
CPCG06V10/26G06F18/23
Inventor 崔晓红肖继海李丹丹相洁李海芳陈俊杰
Owner TAIYUAN UNIV OF 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