摘要
通过抓取网络支付数据包,提取网络交易信息链特征,采用迭代算法搭建网络支付交易多分类器,实现数据分类,并利用机器学习方法构建网络支付欺诈交易动态识别模型.实验结果表明:采用动态识别模型进行网络支付欺诈交易所需时间为12.71 s,传统动态识别所需识别时间为22.63 s,与传统动态识别模型相比,动态识别模型更能节省时间成本,提高了网络支付欺诈交易识别效率.
Research on the dynamic recognition model of online payment fraud transactions based on machine learning is undertaken by capturing network payment data packets,extracting the characteristics of the network transaction information chain,using iterative algorithms to build a network payment transaction multi-classifier,realizing data classification,and using machine learning methods so as to build a dynamic recognition model for online payment fraud transactions.The experimental results show that the time required for online payment fraud transactions using the dynamic recognition model in the project is 12.71 s,and the recognition time required for traditional dynamic recognition is 22.63 s.Compared with the traditional dynamic identification model,the dynamic identification model can save time and cost and improve the efficiency of online payment fraudulent transaction identification.
作者
熊传文
XIONG Chuanwen(College of Computer Science and Information Engineering,Xiamen Institute of Technology,Xiamen 361021,Fujian,China)
出处
《惠州学院学报》
2021年第6期80-84,共5页
Journal of Huizhou University
关键词
机器学习
网络支付
欺诈交易
动态识别模型
数据包
machine learning
network payment
fraudulent transaction
dynamic recognition model
package
