Brain network data enhancement method based on forest auto-encoder

A self-encoder and brain network technology, applied in the field of machine learning data enhancement method theory and application research, can solve problems such as large noise connection, difficult to remove connection and noise at the same time, sparse brain network, etc., achieve fewer parameters and improve classification Performance, the effect of broad application prospects

Active Publication Date: 2020-02-07
BEIJING UNIV OF TECH
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the functional connection of the brain network should be sparse in nature, and our commonly used brain network data is composed of correlation coefficients between every two brain regions, and there are a lot of noise and unnecessary connections.
However, current data generation methods are difficult to remove these unnecessary connections and noises while obeying the original data distribution

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 data enhancement method based on forest auto-encoder
  • Brain network data enhancement method based on forest auto-encoder
  • Brain network data enhancement method based on forest auto-encoder

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0020] Our method mainly includes three parts, the first is the original data generation and parameter initialization, the second is to use the generator based on the forest autoencoder to generate sparse brain network data, and finally use the sparse brain network data generated by multiple random forest selector pairs. Network data is filtered. The basic structure of the method is as figure 1 As shown, its specific implementation steps are as follows:

[0021] Step (1): Raw data generation and parameter initialization, the specific steps are as follows:

[0022] Step (1.1): Raw data generation: First, use the AAL brain atlas segmentation template to select 90 brain regions located in the cerebral cortex in the fMRI data as regions of interests (ROIs), and then calculate the neural network between each two ROIs. Statistical correlation of activity signals, the main measurement methods include Pearson correlation and partial correlation, the adjacency matrix of 90*90 formed ...

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 relates to a brain network data enhancement method based on a forest auto-encoder, and belongs to the field of machine learning data enhancement method theory and application research. The method specifically comprises the following steps of generating original data, generating sparse brain network data and screening the generated sparse brain network data. The sparse brain network data is generated by utilizing the characteristic that the generator based on the forest auto-encoder can influence the sparse degree of decoded data by adjusting the number of trees, and the generateddata is screened by utilizing the screener based on a plurality of random forests, so that the sparsity and robustness of the generated data are ensured. The screened data is used for data enhancement, so that the classification performance of various classifiers is improved. The method has few parameters, the generated data is more consistent with the essential characteristics of the brain network, and the method has a very wide application prospect.

Description

technical field [0001] The invention belongs to the theory and applied research field of machine learning data enhancement method, and specifically relates to a brain network data enhancement method based on a forest autoencoder. Background technique [0002] The human brain is an extremely complex information processing system that can perform complex tasks through the interconnection of multiple neurons, groups of neurons, and brain regions. At the same time, human cognitive processes depend on the interaction of different regions of the brain, and the patterns of these interactions are called functional brain networks. In recent years, more and more studies have shown that many neurological and psychiatric diseases, including Alzheimer's disease (AD), schizophrenia, and autism spectrum disorder (ASD) Usually accompanied by disruption or abnormal integration of connections between brain regions. Therefore, the study of brain networks provides an opportunity to explore th...

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
CPCG06F18/2411
Inventor 王子涵冀俊忠李俊伟
Owner BEIJING UNIV OF TECH
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