Produced by Andy Chase of the indie-pop band Ivy, the record pairs Ms. Hatfield's compact songs with an unabashedly commercial sound, a strategy that works about half the time.
They are aimed at reducing the number of record pairs to be compared in the matching process by removing obvious non-matching pairs without compromising privacy.
Private blocking contributes to removing obvious non-matching pairs and generating candidate record pairs without revealing any information about the originating plaintexts, which reduces the complexity of comparisons.
After generating k-anonymous blocks and corresponding RVs, we need to decide candidate blocks to eliminate record pairs that are expected to be non-matches.
The last step PPRL conducts on each candidate record pairs individually by using a private matching technique, which should not reveal any information regarding the sensitive attributes and non-matches (this step is outside of our approach).
Private blocking [3] aims to generate candidate record pairs which are remained to perform PPRL without revealing any sensitive information that can be used to infer individual records and their attribute values.
Candidate Blocks Generation Given the blocks of each database, through measuring the similarity among the RVs, we can decide whether the records in multiple blocks compare; then, the candidate record pairs would be generated.
The pseudocode of reducing is shown as Algorithm 2. Each reducer traverses all the record pairs which have the same timeWindow id to check whether their distance is within the communication range.
Specifically, suppose c is the number of candidate record pairs produced by the private blocking, (c_m) is the number of true matches among c candidate pairs, n = (|D_A| |D_B|) is the number of all possible pairs, and (n_m) is the number of true matches among all pairs.
Record pairs are then declared as matches or non-matches based on these match scores.
Edith also receives record pairs for comparisons in which one or both values were missing.
