Nonlinear Granger causality detection method based on kernel recursion maximum cross-correlation entropy algorithm
A detection method and kernel recursive technology, applied in the field of time series analysis, can solve the problems that cannot be used to test the causality of time series, poor accuracy of linear regression, wrong causality, etc., to achieve wide application value, reduce impact, and important research meaning effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment Construction
[0022] Below in conjunction with accompanying drawing, the present invention is described in further detail:
[0023] NGC-KRMC uses KRMC as a nonlinear regression model. Compared with traditional GC, GC-RLS and NGC-KRLS, NGC-KRMC not only can correctly detect the nonlinear causal relationship, but also has stronger resolution ability.
[0024] The specific steps of NGC-KRMC are as follows:
[0025] 1) Parameter setting: the maximum order d of the model max , regularization factor λ, Gaussian kernel width σ of MCC 1 , forgetting factor ρ and Gaussian kernel width σ 2 (used by nuclear techniques);
[0026] 2) Using KRMC to establish the autoregressive model and vector autoregressive model of time series X and Y under different model orders;
[0027] 3) Use BIC to select the appropriate order for each model;
[0028] 4) Calculate the causality index by using nonlinear autoregressive error and nonlinear vector autoregressive error;
[0029] 5) Calculate the resolution index...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
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