Time Series Difference Compression for Lossless Wide-Range Data

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for compressing time series data are not suitable for lossless compression, particularly for datasets with a wide range of values, as they either require obtaining original raw data or are not applicable to time series data, leading to inefficiencies in storage and transmission.

Innovation Solution

The apparatus generates difference data by comparing reference data with actual measurement data, reducing the data volume, and then applies lossless compression techniques like run length encoding or dictionary encoding to the difference data for efficient storage and transmission.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If lossless compression is applied to time series data with wide value ranges, then data integrity is maintained, but compression effectiveness is poor

Engineering Contradiction:
Improvedata integrityVSAvoidcompression effectiveness
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent transforms the data representation by calculating differences between consecutive time series values. This parameter transformation converts absolute values with wide ranges into difference values with smaller ranges, enabling effective lossless compression while maintaining the ability to reconstruct original data through reverse calculation.

Inventive Principle:
Principle #35Parameter changes

2Device complexity

If conventional compression methods are used on time series data, then processing is simpler, but compression ratio is insufficient

Engineering Contradiction:
Improveprocessing simplicityVSAvoiddata volume
Core Design Contradiction:
Device complexityVSQuantity of substance

Solution Approach 1:

The patent applies a preliminary transformation step before compression by calculating difference values between consecutive time series points. This preprocessing action creates a transformed dataset with better compression characteristics, allowing conventional lossless compression algorithms to achieve superior compression ratios without increasing operational complexity.

Inventive Principle:
Principle #10Preliminary action

3Volume of stationary object

If difference data is generated and compressed, then storage volume is reduced, but computational cost increases

Engineering Contradiction:
Improvestorage volumeVSAvoidcomputational cost
Core Design Contradiction:
Volume of stationary objectVSPower

Solution Approach 1:

The patent transforms the data representation by calculating differences between consecutive time series values. This parameter transformation converts absolute values with wide ranges into difference values with smaller ranges, enabling effective lossless compression while maintaining the ability to reconstruct original data through reverse calculation.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4404468A1Apparatus for compressing time series data, method for compressing time series data and computer program product
Publication Date: 2024.07.24 HITACHI LTD
  • EP4404468A1 patent drawingFigure 1
  • EP4404468A1 patent drawingFigure 2
  • EP4404468A1 patent drawingFigure 3

AI summary

The present disclosure relates to an apparatus for compressing time series data comprising: a difference detection unit 30 configured to determine difference data from reference data and measurement data, a data compression unit 40 configured to compress the difference data determined by the difference detection unit 30, a data export 50 unit configured to provide the compressed difference data to a data storage device.