Utilizing Linked Data Structures for Social-aware Search Applications

Abstract

Improving the user experience and conversion rate by means of personalization is of major importance for modern e-commerce applications. Several publications in the past have already dealt with the topic of adaptive search result ranking and appropriate ranking metrics. Newer approaches also took personalized ranking attributes of a connected Social Web platform into account to form so called Social Commerce Applications. However, these approaches were often limited to data silos of closed-platform data providers and none of the contributions discussed the benefits of Linked Data in building social-aware e-commerce applications so far. Therefore, we present a first formalization of a scoring model for a social-aware search approach that takes user interaction from multiple social networks into account. In contrast to other existing solutions, our approach focuses on a Linked Data information management in order to easily combine social data from different social networks. We analyze the possible influence of friend activities to the relevance of a person’s search intent and how it can be combined with other ranking factors in a formalized scoring model. As a result, we implement a first demonstrator built upon RDF data to show how an application can present the user an adaptive search result list depending on the users’ current social context.