Abstract
NOVELTY - The method involves selecting an initial attribute for each of two data sources, and selecting a candidate matching function relating to a possible mapping between data in the data sources. The candidate matching function is applied to data relating to the selected attribute in one of the data sources to create possible data relating to the selected attribute in the other data source. A measure of uncertainty relating to an amount of disparity between the sources is derived from the possible data and actual data in the latter source. USE - Used for operating a database management system to integrate data from a data source (claimed). ADVANTAGE - The method accurately matches the attributes and minimizes an overall uncertainty, and is robust against missing information, noisy or erroneous data, and incomplete overlap between sets of real-world entities. DETAILED DESCRIPTION - An INDEPENDENT CLAIM is also included for a database management system for integrating a data from a set of data sources.
Original language | Undefined/Unknown |
---|---|
Patent number | US 07743078EP1708099A1 US20090234869 US7743078 |
Publication status | Published - 2006 |