Reusing calculated match results for MapReduce: Hadoop's MapReduce - Softcover

Inhaltsangabe

The aim of this project is to extend an existing project, called Dedoop, with additional features. Dedoop enables similarities to be identified between pairs of entities within large datasets. The data is processed using MapReduce-based methods. Identifying similarities is a computation that requires a great deal of processing time. If a change has been made to one or more entities in the original set, the calculation must be restarted. The aim of this thesis is to avoid having to repeat the comparisons in full and to find a way, using previously calculated match results, to limit the process to only those specific subsets that require recalculation.

Die Inhaltsangabe kann sich auf eine andere Ausgabe dieses Titels beziehen.