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

Big data filtering method based on multi-core fusion

A technology of multi-core fusion and big data, applied in the direction of impedance network, adaptive network, electrical components, etc., can solve problems such as performance limitations, and achieve good convergence speed and excellent estimation performance

Active Publication Date: 2020-07-14
BEIHANG UNIV
View PDF5 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The performance of the algorithm exhibits limitations when the complexity of the ambient noise increases

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
  • Big data filtering method based on multi-core fusion
  • Big data filtering method based on multi-core fusion
  • Big data filtering method based on multi-core fusion

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0050] The technical solutions of the present invention will be further described below through the embodiments and accompanying drawings.

[0051] see figure 2 with image 3 , a big data filtering method based on multi-core fusion, the steps are as follows:

[0052] Step 1: For each filter initialization, set the parameters of each kernel adaptive filter separately, including kernel width σ, regularization factor γ, and initial fusion coefficient α.

[0053] Step 2: Update the parameter data of each filter according to the single-core adaptive filtering algorithm, and calculate the weight coefficient a at the i-th moment i .

[0054] Step 3: Update the fusion coefficient while training the filter weight coefficient. The fusion coefficient of the m-th filter is the posterior probability of the kernel width when the weight coefficient is calculated as valid. The fusion coefficient must meet the threshold limit.

[0055] Step 4: Perform weighted fusion of multiple independe...

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 big data filtering method based on multi-core fusion, which comprises the following steps: for initialization of each filter, independently setting parameters of each core adaptive filter, including a core width sigma, a regularization factor gamma and an initial fusion coefficient alpha; updating data of parameters of each filter according to a mononuclear adaptive filtering algorithm, and calculating a weight coefficient a _ i at the ith moment; updating the fusion coefficient while the filter weight coefficient is trained, wherein the fusion coefficient of the mthfilter is the posterior probability of the kernel width under the condition that the weight coefficient is calculated to be effective, and the fusion coefficient must meet the limitation of the threshold value; performing weighted fusion of the plurality of independent kernel adaptive filtering according to the fusion coefficient obtained by learning to obtain a final multi-kernel model; and whento-be-predicted input data arrives, performing prediction through the fused multi-core model. The method shows excellent estimation performance in a non-Gaussian environment with complex noise, and has a good convergence rate.

Description

technical field [0001] The invention relates to the field of kernel self-adaptive filtering, in particular to a large data filtering method based on multi-kernel fusion. Background technique [0002] Adaptive filters can automatically adjust the free parameters according to the statistical variation in the environment. Its action principle is to adjust its own response by estimating the statistical characteristics of the input signal, so that a certain cost function reaches the minimum value, usually available figure 1 to determine the auxiliary signal source shown. The auxiliary signal input d(n) can be defined as the expected output of the filter. In this case, the task of the adaptive algorithm is to adjust the weight coefficient of the filter so that the output y(n) of the filter is consistent with the expected output d( The difference e(n) between n) reaches the minimum. Due to the limitations of the computing power of linear systems, complex applications in the real...

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): H03H21/00
CPCH03H21/0043H03H21/0012H03H2021/0076H03H2021/0072H03H2021/0087
Inventor 李文玲褚琳刘杨
Owner BEIHANG UNIV
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