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

Feature selection method based on group sparse norm and local learning

A feature selection method and local learning technology, applied in computer parts, character and pattern recognition, instruments, etc., can solve problems such as ignoring the structure of data groups, and achieve good robustness and accuracy

Inactive Publication Date: 2018-05-18
杭州平治信息技术股份有限公司
View PDF0 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, these methods ignore the group structure of the data

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
  • Feature selection method based on group sparse norm and local learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0024] With reference to accompanying drawing, further illustrate the present invention:

[0025] S1: Use a local learning method to infer the label information of the missing label data in different classes;

[0026] S2: Fit the label information obtained by S1, and use the group sparse specification;

[0027] S3: According to the regression coefficient obtained in S2, calculate the score of each feature, and select a set of features with the highest score.

[0028] Further, the method of local learning of S1 inferring missing label information is as follows:

[0029] (2.1) Organize the data set into an N×M matrix X, N is the number of data, M is the number of features; mark the label information matrix Y, the nth row and the cth column of the matrix Y indicate whether the data n is in the class c ;

[0030] (2.2) Construct the k-nearest neighbor graph of the data set and the kernel matrix K of matrix X; where the k-nearest neighbor graph is used to obtain the k nearest ne...

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 feature selection method based on a group sparse norm and local learning. The method is used for selecting key features representing dataset structures in data to relieve a ''dimension disaster''. By the adoption of an unsupervised learning mode, the method can be applicable to a dataset lacking part of tags. According to the method, first, tags of data lacking tags are inferred through local learning; and then some features most capable of distinguishing different tag data globally are found out on a regressive architecture plus the group sparse norm. The method hasthe advantages that a manifold structure of the data is saved through local learning, a group structure of the data is saved through the group sparse norm, therefore, compared with a traditional feature selection method, the features most capable of maintaining a local structure of the original dataset can be selected through combination of local learning and the group sparse norm.

Description

technical field [0001] The invention relates to a calculation method for feature selection, in particular to feature selection under the action of local learning and group sparse norms, and belongs to the field of computer technology. Background technique [0002] Many technical applications today, such as computer vision, pattern recognition, and data mining, deal with ever-increasing dimensions of data. Higher dimensional data contains more information, but also creates more redundancy and noise. As a result, the "curse of dimensionality" arises, and many methods suitable for low-dimensional data are no longer suitable for processing high-latitude data. [0003] At present, the two main dimensionality reduction methods are feature extraction and feature selection. Feature extraction is to use a set of new features to represent the features of the original data set, and feature selection is to select a representative sub-feature set from the original feature set. Therefo...

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
CPCG06F18/2113
Inventor 余可曼王灿吴越
Owner 杭州平治信息技术股份有限公司
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