DO-Swoosh algorithm maps the input data to domain ontology and count the amount of data for each leaf node. Then define the distance between nodes according to the hierarchical relationships reflected by domain ontology and propose a node merging algorithm based on the principles of data balance and nearest merge. At last, perform entity resolution for each group according to the node merging result. DO-Swoosh still keeps the good generality and gets better performance and scalability with the aid of MapReduce.
KF-Swoosh: An Efficient Spark-Based Entity Resolution Algorithm for BigData
