Efficient reverse far neighbors search

Among the variations of similarity search in data engineering, Reverse Farthest Neighbors(RFN) query is one of the newest and is attracting more attention. RFN is likely to find its applications in facility location problem, product evaluation, social network system, and so on. Given an object set O and a query object q, the RFN query retrieves the objects of O, which take q as their farthest neighbor. The result of an RFN is affected significantly by outliers because it compares only the farthest neighbor. In this paper, we propose a general Reverse Far Neighbors Search and develop an efficient algorithm. Experiments on synthetic and real data confirmed the efficiency.