Using Iterative Computation of Connected Graph Components for Post-Entity Resolution Transitive Closure
Although the Kolb (Kolb et al, 2014) algorithm for the iterative computation of connected graph components was specifically designed for execution in the HDFS environment using Map/Reduce, the algorithm can also be employed in a non-distributed computing environment to support post-entity resolution (ER) transitive closure (blocking). Post-ER transitive closure is an important technique for controlling block size in ER processes, especially when using multiple match keys (Talburt, Zhou, 2015). This paper describes an implementation of the Kolb algorithm in Java, and an example of its application to post-ER transitive closure (blocking) using the OYSTER open source ER system.
Paper
Full text
Using Iterative Computation of Connected Graph Components for Post-Entity Resolution Transitive Closure
Semantic Scholar · Computer Science · 2018
Abstract
Although the Kolb (Kolb et al, 2014) algorithm for the iterative computation of connected graph components was specifically designed for execution in the HDFS environment using Map/Reduce, the algorithm can also be employed in a non-distributed computing environment to support post-entity resolution (ER) transitive closure (blocking). Post-ER transitive closure is an important technique for controlling block size in ER processes, especially when using multiple match keys (Talburt, Zhou, 2015). This paper describes an implementation of the Kolb algorithm in Java, and an example of its application to post-ER transitive closure (blocking) using the OYSTER open source ER system.