ADMM-based unbalanced big data distributed classification method

A classification method and big data technology, applied in the optimization field of convex problems, can solve problems such as high time overhead and slow convergence speed, and achieve the effect of improving classification accuracy, speeding up calculation time, and alleviating class imbalance problems.

Pending Publication Date: 2021-11-09
NANJING UNIV OF SCI & TECH
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Existing empirical studies have shown that the algorithm based on distributed ADMM has slow convergence speed and high time overhead, which is an inherent problem and bottleneck of distributed consensus optimization

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
  • ADMM-based unbalanced big data distributed classification method
  • ADMM-based unbalanced big data distributed classification method
  • ADMM-based unbalanced big data distributed classification method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0106] Our method can solve a problem if it can be written as:

[0107] min x,y f(x)+g(y),

[0108] s.t.Ax+By=C,

[0109] Here f(x) and g(y) are both convex functions, and x and y are variables that satisfy a series of linear constraints. This way the dual of this problem can be solved. Its dual problem can be written in the following form:

[0110]

[0111] Here λ is the dual variable and ρ is the penalty coefficient. In particular, if the local variable x is partitionable, then f(x) can be partitioned into small problems stored on multiple machines. So the problem can be rewritten as follows:

[0112]

[0113] s.t.Ax i +By=C,i=1,2,...,n.

[0114] here x i is the model variable of the small problem on machine i, x=(x 1 ,...,x n ), y is a global variable.

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 an unbalanced big data distributed classification method based on an ADMM, and provides a distributed framework based on the ADMM, a distributed classification problem is divided into some small problems, and the small problems can be solved in parallel through scattered resources; on the basis of a distributed framework, an acceleration strategy is adopted, a more suitable unbalanced data classification model is designed, and the time efficiency is improved. Theoretical analysis and experimental results show that compared with other distributed ADMM methods, the method has the advantages that the convergence speed is higher, the training time is saved, and the expandability of distributed classification on unbalanced data is improved.

Description

technical field [0001] The invention relates to an optimization method for a convex problem, in particular to an ADMM-based distributed classification method for unbalanced big data. Background technique [0002] The classification problem is a very important and universal problem faced by human beings. Classifying things correctly helps people understand the world and makes the chaotic real world organized. For example, automatic text classification is to automatically classify a large number of natural language texts according to certain subject categories, which is a very important issue in natural language processing; text classification is mainly used in information retrieval, machine translation, automatic summarization, information filtering, mail classification tasks. The goal of this task is to classify accurately in imbalanced data samples. Imbalanced data means that the number of samples of one class in the data set far exceeds that of other classes, and the ma...

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): G06K9/62G06F17/10
CPCG06F17/10G06F18/2411G06F18/214
Inventor 王慧慧吴昌胜赵林赵庆玲
Owner NANJING UNIV OF SCI & 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

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