K-Nearest Neighbor, KNN is well-known tool in data mining. KNN query returns objects in a data set that are nearest to a query object-q. One of the shortcomings of KNN is that it is asymmetric. That is, the fact that a query point q has a data point p as its nearest neighbor does not imply that p’s nearest neighbor is q. There may be the case in decision support system where we concern on finding influence set for a query object q. For example, opening a new outlet of company A at particular location is more based on the segment of customers of company B ( competitor of A) who are likely to find new outlet more convenient that the location of B. Such segments of customers can loosely refer as influence set and the reverse relationship is addressed by reverse nearest neighbor, RNN.
In this talk, Deepak will go through some of the papers on the implementation of RNN queries.