Enhanced bloom filter using heuristic optimization for spam detection

Abstract

Bloom filter is a space efficient probabilistic data structure that newlinemakes testing the set membership quick and easy It can identify false newlinepositives but not false negatives The chances of a false positive can be made newlinevery low if the hash bitmap is sufficiently large The probability of false newlinepositives increases with the addition of elements in a set newlineFor spam detection weight is attached to each set of elements The newlinespam weight for a word is a measure used to rate the e mail Each word is newlineassigned to a Bloom filter based on its weight The proposed work introduces newlinean enhanced concept in Bloom filter called Bin Bloom Filter BBF newlineBBF contains a number of Bloom filters with different false newlinepositive rates A high cost Bloom filter in a bin is expected to allocate more newlinehash functions to decrease the membership invalidation rate A low cost newlineBloom filter allows higher false positive probability Hence the BBF newlineminimizes the total membership invalidation cost When the false positive newlinerate is the same for all words present in the set the BBF degenerates to the newlineconventional Bloom filter newline