At present, many information sources are available wherever you are. Most of the time, the information needed is spread across several of those information sources. Gathering this information is a tediousand time consuming job, automating this process would assist the user in its task. Integration of the information sources would provide a (possibly virtual) global information source with all information needed.
A database can be seen to contain information about real world objects. In general, one cannot determine with certainty that this information is correct. Some applications, however, call for uncertain data to be stored in the database, but still allowing the database management system to handle all data management. The key to success in uncertain data, is defining a proper semantics to the underlying model. Instead of describing certain real world objects, an uncertain database describes possible truths about, or appearances of real world objects.
Different appearances of the same real world object are called possibilities. Possibilities of different real world objects together form possible worlds. These are certain views on the world, of which at most only one can describe the real world. We define a theoretical model of a probabilistic database in both a relational and XML setting, a storage model using a traditional RDBMS and probabilistic algebraic operators to query the data. For the relational case, we pay special attention to probabilistic aggregate operators.
|Title of host publication||Flexible Databases Supporting Imprecision and Uncertainty|
|Editors||G. Bordogna, G. Psaila|
|Place of Publication||London|
|Number of pages||25|
|Publication status||Published - May 2006|
|Name||Studies in Fuzziness and Soft Computing|