Associating Social Media to e-Merchandise - A Cold Start Commodity Recommendation

Abstract

In recent years, the boundaries between e-commerce and social networking became increasingly blurred. Many e-commerce websites support the mechanism of social login where users can sign up the websites victimization their social network identities like their Facebook or Twitter accounts. Users can also post their contemporary purchased product on microblogs with links to the e-commerce product websites. Throughout this paper, we incline to propose a novel declare a cross-site cold-start product recommendation that aims to advocate product from e-commerce websites to users at social networking sites in “cold start†things, a retardant that has rarely been explored before. A massive challenge may be thanks to leverage knowledge extracted from social networking sites for a cross-site cold-start product recommendation. We tend to propose to use the coupled users across social networking sites and e-commerce websites (user’s global organization agency have social networking accounts and have created purchases on e-commerce websites) as a bridge to map users’ social networking choices to a clear feature illustration for a product recommendation. In specific, we incline to propose learning every users’ and merchandises’ feature representations (called user embedding and merchandise embedding, respectively) from info collected from e-commerce websites victimization continual neural networks, therefore, apply a modified gradient boosting trees methodology to rework users’ social networking choices into user embedding. We incline to develop a feature-based matrix then resolving approach which could leverage the learned user embedding for a cold-start product recommendation. Experimental results on associate degree outsized dataset made of the most important Chinese microblogging service SINA WEIBO and conjointly the biggest Chinese B2C e-commerce website JINGDONG have shown the effectiveness of our planned framework.