Download Now Free registration required
The authors study the complexity of evaluating aggregate queries on probabilistic databases. The motivation is to manage data produced by integration applications, e.g., data from object reconciliation or information extraction. Standard approaches require that the authors eliminate all uncertainty before any querying can begin, which is expensive in both man-hours to perform the integration and in lost revenue due to down time. An alternative approach where data are allowed to be uncertain, but uncertainty is captured using probabilities has attracted renewed interest.
- Format: PDF
- Size: 343.8 KB