[prev] 4 [next]

Signatures

A signature "summarises" the data from one tuple

A tuple consists of n attribute values A1 .. An

A codeword cw(Ai) is

  • a bit-string, m bits long, where k bits are set to 1  (k ≪ m)
  • derived from the value of a single attribute Ai
A tuple descriptor (signature) is built by combining cw(Ai), i=1..n
  • combine by overlaying codewords   (bitwise-OR)
  • aim to have roughly half of the bits set to 1