Towards Portable Shopping Histories: Using GoodRelations to Expose Ownership Information to E-Commerce Sites

摘要

Recommender systems are an important technology component for many e-commerce applications. In short, they are technical means that suggest potentially relevant products and services to the users of a Web site, typically a shop. The recommendations are computed in advance or during the actual visit and use various types of data as input, in particular past purchases and the purchasing behavior of other users with similar preferences. One major problem with recommender systems is that the quality of recommendations depends on the amount, quality, and representativeness of the information about items already owned by the visitor, e.g. from past purchases at that particular shop. For first-time visitors and customers migrating from other merchants, the amount of available information is often too small to generate good recommendations. Today, shopping history data for a single user is fragmented and spread over multiple sites, and cannot be actively exposed by the user to additional shops. In this paper, we propose to use Semantic Web technology, namely GoodRelations and schema.org, to empower e-commerce customers to (1) collect and manage ownership information about products, (2) detect if a shop site is interested in such information in exchange for better recommendations or other incentives, and (3) expose the information to such shop sites directly from their browser. We then sketch how a shop site could use the ownership information to recommend relevant products.