Synthetic Evaluation Methods of E-commerce Product Quality based on Multi-dimensional Information Fusion

LiangJian Pang,JianZhe Yu,Xinsheng Xu

DOI: https://doi.org/10.1145/3443467.3443862

2020-11-06

Abstract:It is difficult for customers to make correct purchasing decisions due to multiple types and large quantities of product-related data as well as false advertising and promotion at e-commerce platforms during their online shopping process. Combing multi-aspects information to evaluate e-commerce products from the perspective of quality is an effective strategy of obtaining an objective and valid assessment result. With these considerations, an evaluation method for e-commerce products by combining information from stores, products and customers three aspects was put forward in this paper. The index of store reputation was constructed based on the descriptions of stores, and the index of product hot was constructed based on the data of reflecting the popularity of products and its sales. On the basis of online customer reviews, the sentiment score of products was obtained based on product feature extraction, clustering and modeling of emotional modified structure. These indexes reflect the quality guarantee and status of e-commerce products from their own independent point of view. Then, a synthetic evaluation model of e-commerce products was established by fusing these indexes. The technical framework of e-commerce product synthetic evaluation was presented as well. Finally, several experiments were performed to verify the feasibility and effectiveness of our proposed methods. It provides a promising method for customers to distinguish e-commerce products and guide their online shopping.