Recording Provenance on Probabilistic Databases

Tracking data provenance (or lineage) has become increasingly important in many large-scale applications. Till now, a few methods have been proposed to record data provenance. However, most of them mainly focus on deterministic databases except Trio style lineage that aims at probabilistic databases. Processing provenance upon probabilistic database is even challenging because of the exponential growth of possible world instances and dependence of intermediate tuples. In this paper, we propose a model, named PWP-tree, to record data provenance upon probabilistic database. We further present some data provenance model upon probabilistic databases, including Trio style lineage, can translate them into our model for uncertainty propagation. Compared with Trio style lineage, our model is independent of intermediate results, but the results of probability evaluation is identical to the possible worlds semantics. Detailed experimental results show the effectiveness, efficiency and scalability of our proposed model.