The rapid development of information technology and digitalization has promoted the spread of e-commerce. The variety of products can make customers feel overwhelmed when choosing and using them. Currently, Customized E-commerce B2C Recommendation Systems bring convenience to both customers and businesses by improving sales efficiency, solving the problem of information overload, and optimizing the supply chain. Amazon is an American multinational technology company that specializes in e-commerce and has been referred to as one of the most influential economic and cultural forces in the world. For this project, I will design a Customized E-commerce B2C Recommendation System for Amazon by analyzing Amazon Review Data (Ni, 2018), which is a collection of reviews including ratings, text, helpfulness votes, product metadata like product descriptions, category information, price, brand, image features, and links that are viewed. All analysis tools are from Amazon Web Services (AWS), such as Amazon Textract, Amazon Comprehend, Amazon A2I, Amazon Kendra, Amazon Personalize, Amazon Translate, etc. (AWS website). There are four stages in this project: retrieval, filtering, scoring, and ordering. In this paper, I will focus on the retrieval stage. The method of the retrieval stage should be based on a hybrid system called Enhanced Augmented Two-Tower Modeling, which combines Collaborative Filtering (CF) and Content-Based Filtering (Yu, Wang, Feng, Xue, 2021). The goal of the project is to build customer profiles and recommend products/services by matching customer preferences and product information based on textual information like product descriptions and customer comments.
Jingwen Ni (Sat,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: