Towards Improving Multi-query Search Approaches

摘要

In this paper, we present our continuous work on similarity search problems. In our previous work [14], we discussed the meaning of K nearest neighbors from a new perspective and proposed an algorithm to find nearest neighbors to multiple queries with weights, which works for both full similarities and partial similarities in subsets of dimensions. In this paper we present the improvement and enhancement of our original approach to locating good candidates for multi-query search.