论文部分内容阅读
本文使用网购平台储存的大量商家商品信息,构建仿冒品识别测量模型。以京东商城NIKE男鞋为例,本文一方面采集了商品网页结构化数据,另一方面也为得到能判断商品真假的基础数据做了大量工作,即用KJ法得到做工、气味、鞋底和鉴定结果作为指标,根据商品买家评论判断商品的真假,基于K2算法对结构化指标和商品真假结果进行贝叶斯网络结构的学习,选出最佳贝叶斯网络结构并学习参数,得到的模型可作为推算出仿冒品概率的依据。
This article uses a large amount of merchant merchandise information stored on the online shopping platform to construct counterfeit goods identification measurement models. Take Jingdong Mall NIKE men’s shoes as an example, on the one hand, this article collects the structured data of the product web page, and on the other hand, it does a lot of work to get the basic data that can judge the true and false of the product, that is, using KJ method to get workmanship, odor, sole and The identification result is used as an indicator to determine whether the product is authentic or not according to the commodity buyer’s review. The Bayesian network structure is learned based on the K2 algorithm for the structured index and the product’s true and false results. The best Bayesian network structure is selected and the parameters are learned. The obtained model can be used as a basis for deducing the probability of counterfeit goods.