基于稳定度的异常商品识别模型研究与实现
Research and Implementation of Abnormal Product Identification Model Based on Stability

随着经济的发展，网购现已得到全方位的普及，因其有方便快捷、省时省力、送货上门等优点，越来越受到人们的青睐，成为日常生活中不可或缺的一部分。随着人们经济能力、消费水平的提高，对网购体验的需求也愈发上涨。同时网上各大零售业务间的竞争也愈发激烈，为了能够吸引消费者的目光以增加商品的销量，某些商家开始采用刷销量、刷好评、删差评等“炒信”“刷单”手段对商品进行恶意推广，侵犯消费者权益。为保障消费者的知情权和选择权，本项目通过浪潮卓数公司提供的数据集，通过数据挖掘定量分析和定性分析结合的方法来剖析商品出现异常的原因，采用数学建模和机器学习的方法，定义部分异常商品指标，并利用这些指标构建出查找和预测异常商品的模型。实验结果表明该模型效果较好，具有一定的实用性。
With the development of the economy, online shopping has gained widespread popularity in all aspects. Due to its advantages such as convenience, speed, time and effort saving, and door-to-door delivery, it is increasingly favored by people and has become an indispensable part of daily life. With the improvement of people’s economic ability and consumption level, the demand for online shopping experience is also increasing. At the same time, competition among major online retail businesses has become increasingly fierce. In order to attract consumers’ attention and increase product sales, some businesses have started to use “speculation” and “order” methods such as selling, positive reviews, and negative reviews to maliciously promote products, infringing on consumers’ rights and interests. To protect consumers’ right to know and choose, this project uses a dataset provided by Inspur Zhuosu Company to analyze the reasons for abnormal products through a com-bination of quantitative and qualitative data mining analysis. Mathematical modeling and machine learning methods are used to define some abnormal product indicators, and these indicators are used to construct a model for finding and predicting abnormal products. The experimental results indicate that the model has good performance and certain practicality.