Industrial internet of things high-frequency data compression method based on time segmentation and clustering

By improving the elephant herd optimization algorithm and Gaussian mixture model to segment and cluster time series data, and combining DPCM and Huffman coding, the problems of low compression rate and insufficient data restoration capability of time series data in the Industrial Internet of Things are solved, achieving more efficient data compression and decompression.

CN115459782BActive Publication Date: 2026-06-26BEIJING YUNSONG ZHITONG TECHNOLOGY CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
BEIJING YUNSONG ZHITONG TECHNOLOGY CO LTD
Filing Date
2022-08-29
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

Existing time series compression methods suffer from low compression ratios and insufficient data restoration capabilities in the Industrial Internet of Things (IIoT). In particular, the SAX method performs poorly when faced with large data variations and lacks methods that utilize the correlation between different time series for compression.

Method used

We employ a Gaussian segmentation model based on an improved elephant herd optimization algorithm to partition the time series, and combine it with an improved peak density-initialized Gaussian mixture model for clustering. We then use the SAX method to transform the cluster centers into symbolic representations, and combine DPCM and Huffman coding to compress timestamps and cluster labels. Finally, we propose a re-segmentation and shearing strategy to improve compression ratio and data restoration capability.

Benefits of technology

It achieves lower compression ratios and better data restoration capabilities, is highly adaptable, and can effectively handle the compression and decompression of high-frequency data in the Industrial Internet of Things.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115459782B_ABST
    Figure CN115459782B_ABST
Patent Text Reader

Abstract

The application discloses a kind of industrial internet of things high-frequency data compression methods based on timing segmentation and clustering, belongs to time series compression technical field, it includes: using the improved Gauss segmentation model based on image group optimization algorithm to time series division, obtain multiple segments;Using the improved peak density initialization based on Gauss mixture model to its clustering, obtain module variance and clustering label;According to the proportion of variance, each clustering center is again equidistant segmentation;Using SAX method, the mean of each segment of each center is converted into symbolic representation, the first character of each class is capital letter;Same symbol representation is cut to the same class, retain the first capital letter, obtain time series value compression data;Obtain timestamp and segmentation point information, using DPCM algorithm to compress it;Using Huffman coding, clustering label is symbolized and compressed.The method can make industrial internet of things data obtain lower compression ratio and better data restoration ability.
Need to check novelty before this filing date? Find Prior Art