Table 3.
Comparative analysis of existing linkage tools.
| Feature | RecLink | PLA | AtyImo | CIDACS-RL | FRIL | Febrl |
|---|---|---|---|---|---|---|
| Deterministic | Pure Comparison | Exact Comparison | Hybrid approach | Exact query | Equality function | Exact comparison functions |
| Probabilistic | Character Sequence and fuzzy | Automatic codes | Fully probabilistic | Semi-exact and fuzzy queries | Edit distance, soundex and Q-gram | Approximate comparison functions |
| Blocking | One step (single attribute) and multi-step predicates) | No | Predicates | TF-IDF indexing | Nested loop join and Sorted neighbourhood | Block, Ssorted and fuzzy (bigram) |
| Anonymization | No | No | Bloom Fliter | No | No | No |
| Manual review of Dubious records | No | PLA-MR | Second round with adjusted cut-offs | Yes | Yes | Yes |
| Automated review of dubious records | No | PLA-FAP | Machine learning-based | No | Yes (expectation maximization) | Expected |
| Open source, freely available | Yes(GPL) | No | Yes | No | Yes | Yes |