摘要
随着比特币等数字货币的快速发展,通过发行代币进行融资的ICO得到广泛关注,利用ICO进行传销诈骗的事件也逐渐涌现。建立类罪模型对ICO进行类型区分、提取犯罪特征是当前经济犯罪侦查取证领域的热点课题。由于ICO分类受到诸多因素的影响,得到精确的分类结果比较困难。收集了国内外150组ICO相关数据,利用R型聚类分析提取ICO指标作为聚类特征,利用Q型聚类分析对ICO数据进行聚类。结果表明,传销型ICO和非传销型ICO满足的指标值有一定的差异,聚类划分准确率达94%以上,与以人工侦查传销犯罪为依据的传统分类有良好的一致性,预测未知ICO数据所属类别的准确率为84%。
With the rapid development of bitcoin and other digital currencies,ICO,which has been financed by issuing encrypted tokens,has received widespread attention,and many incidents of pyramid selling fraud are gradually emerging by ICO.It is a hot spot to establish the type of crime model to distinguish the type of ICO and to extract the characteristics of the crime in the field of economic crime investigation.Since ICO classification is affected by many factors,it is difficult to obtain accurate classification results.Collecting 150 sets of ICO related data at home and abroad,using R-type clustering analysis to extract ICO index as clustering feature,and using Q-type clustering analysis to cluster ICO data.The results show that there is a certain difference between the index values met by the pyramid-based ICO and the non-municipal ICO,and the clustering accuracy rate is over 94%,which is in good agreement with the traditional classification based on human investigation and pyramid schemes.The accuracy rate of prediction the category of unknown ICO data is 84%.
作者
刘二根
王露
左黎明
艾美珍
张梦丽
LIU Er-gen;WANG Lu;ZUO Li-ming;AI Mei-zhen;ZHANG Meng-li(School of Science,East China Jiaotong University,Nanchang 330013,China;SEC Institute,East China Jiaotong University,Nanchang 330013,China;Collaborative Innovation Center for Economics Crime Investigation and Prevention Technology,Nanchang 330013,China)
出处
《宜春学院学报》
2018年第12期4-8,108,共6页
Journal of Yichun University
基金
国家自然科学基金项目(11761033)
江西经济犯罪侦查与防控技术协同创新中心开放基金资助课题(JXJZXTCX-001)
江西省教育厅科技项目(GJJ161417
GJJ170386)
江西省交通运输厅科技项目(2017D0037)
关键词
ICO
R型聚类分析
Q型聚类分析
传销
ICO
R-type clustering analysis
Q-type clustering analysis
pyramid selling
