A New Distributed Name Disambiguation System Based on MapReduce

摘要

Social network search is a kind of vertical search based on the information polymerization of social network data.And name disambiguation is one of the vital issues in social network search.With the explosive growth of the information,effectively deal with name disambiguation in massive data scenario becomes an important issue.To tackle this issue,we combine the MapReduce model and document clustering algorithm to propose a distributed method for name disambiguation.And then present a distributed name disambiguation system.This system runs on the Hadoop platform,and makes the name disambiguation parallelized by dividing the document clustering task to a couple of maps and reduces.In addition,we evaluated our system in terms of expansibility and accuracy.