基于蜂窝网络实测数据的电信诈骗检测模型研究
首发时间:2019-09-23
摘要:随着无线通信技术的迅猛发展,利用手机电话等通讯手段进行电信诈骗,已经成为危害用户安全的一大问题。针对这一问题,论文通过对蜂窝网络数据进行深入分析与挖掘,构建电信诈骗用户检测模型,该模型包括数据预处理,CNNcombine算法、模型评估三部分。首先在数据处理部分,对数据集进行了特征筛选、编码、抽样等工作。其次,CNNcombine算法是本文提出的一种一维卷积神经网络与多个传统分类算法结合的算法,将卷积神经网络运用于解决文本像信号以外的分类问题。最后在模型评估部分,证明了CNNcombine算法比XGBoost等常用机器学习分类算法检测电信诈骗用户上有更高的准确率。
For information in English, please click here
Research on Telecom Fraud Detection Model Based on Cellular Network Data
Abstract:With the rapid development of wireless communication technology, the use of mobile phones and other means of communication for telecommunications fraud has become a major problem that endangers user security. Aiming at this problem, this paper constructs a telecom fraud user detection model by in-depth analysis and mining of cellular network data. The model includes data processing, CNNcombine algorithm and model evaluation. First, in the data processing part, the data set is subjected to feature screening, coding, sampling, and the like. Secondly, the CNNcombine algorithm is a combination of a one-dimensional convolutional neural network and multiple traditional classification algorithms. The convolutional neural network is applied to solve classification problems other than text image signals. Finally, in the model evaluation part, it is proved that the CNNcombine algorithm has higher accuracy than the common machine learning classification algorithm such as XGBoost to detect telecom fraud users.
Keywords: Machine learning cellular network data deep learning classification algorithm
基金:
引用
No.****
同行评议
共计0人参与
勘误表
基于蜂窝网络实测数据的电信诈骗检测模型研究
评论
全部评论0/1000