摘要
随着区块链技术的发展,以比特币为代表的链上交易量迅速增长,隐藏在其中的非法交易严重影响链上交易。当前区块链异常交易检测中有逻辑回归和随机森林方法,与逻辑回归相比,随机森林准确度较高但存在训练耗时较长的问题,本文在分析逻辑回归和随机森林的基础上提出了并行随机森林模型,实验表明并行随机森林在保持准确性的前提下,大幅减少了训练时间。
With the development of blockchain technology,the volume of transactions on the chain represented by Bitcoin has grown rapidly,and illegal transactions hidden in it have seriously affected the in-depth de-velopment of transactions on the chain.The current blockchain abnormal transaction detection methods include logistic regression and random forest,compared with logistic regression,the random forest has higher accuracy but has the problem of longer training time.This thesis proposes parallel random forest model based on the analysis of logistic regression and random forest model.The experiment shows that under the permise of maintaining accuracy for parallel random forest model,greatly ruducing train time.
作者
赵永斌
陈苗
李涛
尤军考
ZHAO Yong-bin;CHEN Miao;LI Tao;YOU Jun-kao(School of Information Science and Technology, Shijiazhuang Tiedao University,Shijiazhuang Hebei 050043,China;China Railway Beijing Bureau Group Co., Ltd. Shijiazhuang Electricity Services Department,Shijiazhuang Hebei 050000,China;China Mobile Group Hebei Company Limited, Shijiazhuang Hebei 050000,China)
出处
《河北省科学院学报》
CAS
2021年第5期1-8,共8页
Journal of The Hebei Academy of Sciences
基金
河北省研究生示范课程建设项目(KCJSX2020062)
河北省高等学校科学研究计划课题重点项目(ZD2020174)。
关键词
区块链
异常交易检测
逻辑回归
随机森林
并行计算
Blockchain
Abnormal transaction detection
Logistic regression
Random forest
Parallel computation
