When looking inside a marketplace, trust and provenance tracking between all participants is the key to establish commercial relationships. Especially in the LOD cloud, trust is an essential mechanism due to the fact that nearly everybody is (potentially) able to manipulate information or their interlinking concepts. Moreover, such an open approach also challenges the development of proper value chain mechanisms.
The Linked Open Data initiative aims to foster data provenance information to ensure high quality of retrievable data. In contrast to that, a trust layer based on social interactions between participants still plays a minor role. Following this, CODE will research techniques to establish a trust layer exploiting information available in the market place:
- user profiles
- history of previous transactions
- social relations between users
User profiles inherit lots of information, e.g., research interest or cited articles, which can be used to recommend users with similar affection. Besides exploration of user profiles, an essential part is to investigate the history of yet ended transactions and to gain knowledge of which users have been involved. Finally, existent social connections play a crucial role in the creation of trust, e.g. shortest path between two friendships. These disparate information points will be consolidated and combined with data provenance to ensure a reasoned value chain.
We will research and develop algorithms and mechanisms to establish a trust and provenance based value chain supporting different user roles and allowing the evolution of such roles. This will be accomplished through the following topics:
- Deriving trust information between users involving different dimensions (user profiles, history, content).
- Combining trust information between users with data provenance to track potential revenue streams.
- Development of value chain mechanisms and participant roles based on trust and provenance mechanisms.
While value chains and participant roles in auctioning and work crowd-sourcing are clear, a detailed analysis and empirical investigation for data marketplaces is still missing. Moreover, value chains in data marketplaces, especially those considering data analysts and integrators, remain unclear. Potential scenarios include, but are not limited to, auctions over data, traditional revenue models or crowd-sourcing of analytical & integration work tasks similar to Amazon’s Mechanical Turk.