Accurate Multi-Dimension Counting Bloom Filter for Big Data Processing
Li We · Dianzi xuebao · 2015
Based on the analysis of multi-dimension Bloom filter presented before,an accurate multi-dimension counting Bloom filter( AMD-CBF) query algorithm based on bijective function is proposed for big data processing. When representing or querying an element,AMD-CBF needs two steps. The first step is to hash and map each attribute of the element to their corresponding accurate counting Bloom filter( A-CBF); The second step is to transform all attributes of the element into a value by bijective function to represent the overall information of the element,then the value is hashed and mapped into a combined counting Bloom filter( C-CBF) for completing the representation and query confirmation of the elements overall. Both theoretical analysis and experiment showthat the AMD-CBF can support concise representation,approximate membership query and deletion of multidimension data set and significantly lower false positive rate and improve query accuracy compared to similar research.