The dynamic Bloom Filter with compressed

Bloom Filter is an efficient data structure for membership queries on a set with a certain known false positive probability. The traditional Bloom Filter only can represent static set and it just focuses on how to decrease the false positive probability. But in practice more applications use Bloom Filter to represent dynamic set than to represent static set. To address this issue, this paper propose a novel Bloom Filter to represent dynamic set and design necessary operations such as item insertion, item deletion, membership query and filter union. As is well-known, in order to support delete item, Counting Bloom Filter (CBF) have to be used. The proposed Bloom Filter also uses CBF but we introduce the idea of a hierarchical structure to improve the performance of CBF. The target could be the implementation of the compressed data structure in the small memory with decreasing the false positive probability.