Attractor Reconstruction for Time-Series Feature Extraction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing data analysis techniques using topological data analysis (TDA) struggle to accurately extract features from time-series data due to limited data points, leading to deteriorated feature extraction performance.

Innovation Solution

The method involves reconstructing an attractor using time-series data with interpolated points at regular intervals, ensuring a consistent number of data points per timing, which stabilizes the Betti series and enhances feature extraction by increasing the density and clarity of the attractor shape in the phase space.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the number of data points in time-series data is limited, then the data analysis process is simpler and faster, but the feature extraction performance deteriorates

Engineering Contradiction:
Improvefeature extraction accuracyVSAvoidnumber of data points
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent segments the time-series data by dividing each time point into multiple interpolated points at regular intervals. This segmentation increases the number of data points without adding new information, allowing the attractor to be reconstructed with higher density while maintaining the same underlying data content.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary interpolation of the time-series data before constructing the attractor. By pre-interpolating the data at regular intervals, the system prepares a denser dataset that stabilizes the Betti series and improves feature extraction accuracy before the actual analysis begins.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If the number of data points is increased through interpolation, then the attractor shape becomes clearer and Betti series stabilizes, but the data processing time increases

Engineering Contradiction:
Improveattractor shape clarityVSAvoiddata processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies periodic interpolation at regular intervals between data points. This periodic action creates a consistent pattern of interpolated points that stabilizes the Betti series and clarifies the attractor shape. The regularity of the interpolation pattern allows for efficient processing while achieving the desired enhancement in attractor clarity.

Inventive Principle:
Principle #19Periodic action

Data Source

PatentEP3923228B1Data analysis method, device and program
Publication Date: 2023.08.16 FUJITSU LTD
  • EP3923228B1 patent drawingFigure 1
  • EP3923228B1 patent drawingFigure 2
  • EP3923228B1 patent drawingFigure 3

AI summary

A program causes a computer to execute a process, the process including determining numerical values indicating features at respective timings having a predetermined time interval with respect to time-series data to be analyzed, numbers of the numerical values at the respective timings being made same, and generating an attractor related to the time-series data based on the determined numerical values.