3 ms·

Not sure I follow. For detecting uniques all you care about is the definite negatives? Either way, the author addresses precision in his post.

• reitzensteinm 7y ago
  False positives mean you're throwing away genuinely unique items that hash the same. If you use a 256 bit Bloom filter to process a billion items, you'll get exactly 256 results that are indeed unique. The rate of false positives you require out of the data structure is key. If your program is correct with a 20% false positive rate, you're golden. If the goal is more or less 0, look elsewhere. The author addresses precision, but not in a way that questions whether a Bloom filter is indeed the right tool for the job.
  • jiofih 7y ago
    But you don’t care about false positives, only the definite negatives! I was trying to address the fact that the top comment on the utility of a bloom filter is covered in the post.
    • reitzensteinm 7y ago
      If you don't care about the false positives, you don't need a Bloom filter. Just reject everything.