基于非平衡数据的深度再分片算法

打开文本图片集

中图分类号：TP309.2；TP393 文献标识码：A 文章编号：2096-4706（2025）07-0114-06

Abstract： With the popularization of blockchain technology，the surge intransaction volume poses chalenges to network performance.Traditional segmentation methods based on account transactions perform poorly when dealing with unbalanced data，which leads to an increase incross-segmentation transactionsand affects performance.Inresponse tothis situation， aDeep TraversalRe-segmentationAlgorithmbasedon transaction frequencyis proposed.Thisalgorithmreduces crosssegmentationtransactionsthroughinitialsegmentationandtenconductsdep traversalre-segmentationonthesegmentation withdense transactionstosolvetheproblemofdataimbalance.Bycalculating thetransaction frequencyofaccounts，itensures load balancingandattempts to allcate accounts with frequent transactions to the same segmentationas much as possible. Experimentalresultsshowthatthisalgorithmcanefectivelybalancetheloadreducecro-segmentationtransactions，and improve network throughput and efficiency.

Keywords： blockchain technology; Deep Traversal Re-segmentation Algorithm; transaction frequency; crosssegmentation transaction

0 引言

区块链技术源自中本聪于2008年提出的比特币系统，区块链技术自诞生以来，以其去中心化、安全性高、数据不可篡改等特点，被广泛应用于金融、供应链、物联网等多个领域。（剩余7592字）