My questions is: What is the "real" difference between Reverse nearest neighbor(RNN) and Nearest neighbor(NN) queries? In my opinion, these two should return same result.


The definition here states that

Given a multi-dimensional dataset $P$ and a point $q$, a reverse nearest neighbor (RNN) query retrieves all the points $p\in{P}$ that have $q$ as their nearest neighbor.

As a simple example, consider the 1D point set $P=[a,b,c]=[0,2,3]$. Then we have \begin{align} \mathrm{NN}(a) &= b \\ \mathrm{NN}(b) &= c \\ \mathrm{NN}(c) &= b \end{align} So \begin{align} \mathrm{RNN}(a) &= \emptyset \\ \mathrm{RNN}(b) &= \{a,c\} \\ \mathrm{RNN}(c) &= b \end{align} Does this help?

  • $\begingroup$ I got the logic, Exactly what i was looking for simple yet informative. Thank you! $\endgroup$ – user331172 May 29 '17 at 4:21
  • $\begingroup$ @GeoMatt22 The reverse nearest neighbour relation is not a symmetric relation. $\endgroup$ – Dylan May 29 at 13:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.