基于深度学习的商品价格欺诈行为与电子数据取证关联应用

Affiliate Application of Commodity Price Gouging and Digital Forensics Based on Deep Learning

该研究聚焦电商领域商品价格欺诈行为的智能检测问题,采用深度学习技术结合电子数据取证手段构建商品价格监测系统。研究团队使用爬虫从各大电商平台抓取商品价格数据,并实施严格的数据清洗与质量控制,选取LSTM网络架构,利用其在处理时间序列数据的优势,捕捉商品价格变动规律。通过嵌入层对离散特征编码,多层LSTM单元捕获时序特征,全连接层用于输出价格预测。电子数据取证在此过程中起到关键作用,能揭示商品价格欺诈的具体模式,大大提高了欺诈案件的侦破效率。尽管现有模型和方法已取得一定成效,但研究也有局限性,如数据代表性不足、模型泛化能力、标签不平衡问题及模型可解释性差等。

This paper focuses on the intelligent detection of commodity price gouging in the e-commercial field,and uses deep learning technology and digital forensics methods to build a commodity price monitoring system.The research team used web crawlers to extract commodity price data from major e-commercial platforms.Through strict data cleaning and quality control,the team selected the LSTM network architecture and leveraged its advantages of processing time series data to capture the regularity of commodity price changes.Discrete features are encoded through the embedding layer,multi-layer LSTM cells capture time-order features,and fully connected layers output price predictions.Digital forensics plays a key role in this process,revealing specific patterns of commodity price gauging and greatly improving the detection efficiency of fraud cases.Although existing models and methods have achieved certain results,the research has some limitations,such as insufficient representativeness of data,problems of model generalization ability and label imbalance,and poor model interpretability.