Entity-aware query processing for heterogeneous data with uncertainty and correlations

Many modern systems rely on rich heterogeneous data that has been integrated from a variety of different applications and sources. To successfully perform their tasks, these systems require to know which data refer to the same real-world entities, such as locations, people, or movies. My work focuses on addressing this requirement through a new approach for entity-aware query processing over heterogeneous data. Data provided for integration is processed to generate the possible entities and linkages between these entities. This information is never merged with the original data, but used during query processing to provide entity-aware results that reflect the real-world entities existing in the current data. Special emphasis is given to the effective management of uncertainty and correlations that either exist in the original data, or are generated by data matching techniques.