Predicting Critical Transitions via Associative Transfer Entropy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods fail to effectively detect and predict critical transitions in heterogeneously networked dynamical systems, particularly due to their inability to analyze directional influences and dynamics changes over time.
Innovation Solution
A system that transforms multivariate time series into symbolic series, calculates pair-wise transfer entropy, and determines associative transfer entropy measures to predict critical transitions by estimating trajectories using the natural logarithm and discrete time steps, allowing for the analysis of directional influences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If transfer entropy is used to analyze financial market data, then asymmetric influence can be detected, but the method cannot handle dynamics and structure changes in directional influence
Solution Approach 1:
The patent implements a dynamic sliding window approach that continuously updates transfer entropy calculations over time. The system divides time series data into overlapping windows and recalculates transfer entropy for each window, enabling detection of temporal changes in directional influence. This dynamic implementation resolves the contradiction by making the measurement adaptive to changing system conditions while maintaining precision through localized analysis.
Solution Approach 2:
The system performs preliminary calculations of transfer entropy across multiple time windows before identifying critical transitions. By pre-processing the data to establish baseline directional influences and their temporal evolution, the system prepares the foundation for detecting abrupt changes. This preliminary action enables the system to distinguish between normal dynamics and critical transitions more effectively.
2Measurement precision
If spectral early warning signals theory is used to detect critical transitions, then system structure and network connectivity can be estimated, but the symmetric nature of covariance spectrum does not permit analysis of directional influences
Solution Approach 1:
The patent explicitly introduces asymmetry by using transfer entropy instead of symmetric covariance measures. Transfer entropy is inherently directional, allowing the system to distinguish information flow from source to destination. This asymmetric measurement preserves directional influence information while still enabling estimation of system structure and network connectivity through the pattern of directional dependencies across multiple variables.
3Productivity
If symbolic transfer entropy is used to detect asymmetric dependences, then computational efficiency is improved, but the emphasis is not on identifying the dynamics of the complex system structure
Solution Approach 1:
The patent combines symbolic transfer entropy with dynamic sliding window analysis to simultaneously achieve computational efficiency and capture system dynamics. The symbolic approach maintains efficiency by discretizing continuous data, while the temporal evolution of symbolic patterns across windows reveals dynamic structural changes. This combination resolves the contradiction by making the efficient symbolic method sensitive to temporal dynamics.
Solution Approach 2:
The system adds the temporal dimension by analyzing how symbolic transfer entropy patterns evolve across multiple time windows. This transformation from static to dynamic analysis allows the efficient symbolic method to capture system structure dynamics by examining the evolution of asymmetric dependencies over time, effectively adding a temporal dimension to the analysis.
Data Source
AI summary
Described is a system for predicting system trajectories toward critical transitions. The system transforms a set of multivariate time series of observables of a complex system into a set of symbolic multivariate time series. Then pair-wise time series of a transfer entropy (TE) measure are determined, wherein the TE measure quantifies the amount of information transfer from a source to a destination in the complex system. An associative transfer entropy (ATE) measure is determined which decomposes the pair-wise time series of TE to associative states of asymmetric, directional information flows, wherein the ATE measure is comprised of an ATE+ influence class and a ATE− influence class. The system estimates ATE+, TE, and ATE− trajectories over time, and at least one of the ATE+, TE, and ATE− trajectories is used to predict a critical transition in the complex system.