PROJECT TITLE:

Connecting Social Media to E-Commerce Cold-Start Product Recommendation using Microblogging Information - 2016

ABSTRACT:

In recent years, the boundaries between e-commerce and social networking have become increasingly blurred. Many e-commerce Internet sites support the mechanism of social login where users can sign up the Web sites using their social network identities like their Facebook or Twitter accounts. Users will also post their newly purchased merchandise on microblogs with links to the e-commerce product Web pages. In this paper, we have a tendency to propose a unique solution for cross-website cold-begin product recommendation, which aims to suggest products from e-commerce Net sites to users at social networking sites in “cold-start” situations, a downside which has rarely been explored before. A major challenge is a way to leverage knowledge extracted from social networking sites for cross-web site cold-start product recommendation. We have a tendency to propose to use the linked users across social networking sites and e-commerce Internet sites (users who have social networking accounts and have made purchases on e-commerce Internet sites) as a bridge to map users' social networking features to a different feature representation for product recommendation. In specific, we tend to propose learning both users' and product' feature representations (referred to as user embeddings and products embeddings, respectively) from knowledge collected from e-commerce Internet sites using recurrent neural networks and then apply a modified gradient boosting trees methodology to rework users' social networking options into user embeddings. We then develop a feature-primarily based matrix factorization approach which will leverage the learnt user embeddings for cold-start product recommendation. Experimental results on a massive dataset created from the largest Chinese microblogging service Sina Weibo and the most important Chinese B2C e-commerce website JingDong have shown the effectiveness of our proposed framework.