Abstract：To solve the problems of low query efficiency and multiple access to objects in NRA algorithm and BPA algorithms, an efficient topk query algorithm (TKBFP) was proposed based on three-dimensional bracket bloom filter (TF). TF was used for data processing in the algorithm to realize low false positive rate and access efficiency with reduced memory consumption. The optimal location index strategy was adopted to avoid multiple access to data objects. A strict semantic definition of TKBFP algorithm was given to deduce the number of hash function in each BF. The TKBFP algorithm was tested by the simulation platform of independent development java program to evaluate the efficiency and the storage perfor-mance of the algorithm. The results show that the algorithm can effectively avoid multiple access to data objects to obtain efficient query processing for large-scale data with low false positive rate. Compared with NRA and BPA, the proposed algorithm cost is reduced significantly when the attribute list number are more than four. The proposed algorithm is suitable for large-scale data query processing.
2012, Vol. 33 
