基于用户关系与HCM的APP虚假用户识别

APP fake user identification based on user relationships and HCM

随着移动互联网的迅速发展,APP已成为人们生活中不可或缺的一部分。同时,APP应用市场的蓬勃发展也催生APP虚假用户这一市场乱象。在这样的背景下,进行APP虚假用户识别的研究对于开发者获得最真实有效的软件使用反馈,同时帮助软件下载者了解最真实的软件使用体验具有重要的现实意义。文中提出一种基于HCM模型的APP虚假用户识别算法。首先,根据APP用户数据构建用户关系图;接着,提出一种基于异构图神经网络(HetGNN)与卷积神经网络(CNN)的联合模型来识别APP虚假用户。通过相关算法进行实验,结果证明了该算法的有效性,能够准确识别出APP虚假用户。

With the rapid development of mobile Internet,mobile APPs have become an indispensable part of people's lives.However,the booming development of the APP market has also led to fake APP users.In this context,research on identifying fake APP users is of great practical significance for developers to obtain the most authentic and effective software usage feedback and for software downloaders to understand the most authentic software usage experience.In view of this,an APP fake user identification algorithm based on the HCM(joint HetGNN and CNN⁃based model)is proposed.A user relationship graph is constructed based on the APP user data.A joint model based on heterogeneous graph neural network(HetGNN)and convolutional neural network(CNN)is proposed to identify fake APP users.Experiments were carried out by related algorithms.The experimental results show that the proposed algorithm is of effectiveness and can identify fake APP users accurately.