Doing entity resolution on entity resolution you can find: record linkage, duplicates detection, entity matching, deduplication, fuzzy matching.
Doing entity resolution on records you can find: tuples, entity profiles, semi-structured documents.
It helps to scale Entity Resolution: it efficiently extracts loose schema information, and uses this information to group together records that most likely will match.
Basically, instead of comparing all possible paris of records, you only compare subsets of them.
P.S.: Blast employs only unsupervised techniques.
When you have semi-structured data to clean, but you cannot do schema-matching to apply traditional blocking techniques.
Here the code of "BLAST: a loosely schema-aware meta-blocking approach for entity resolution", please cite: .
The approach is implemented on top of the Blocking Framework, A framework for blocking-based Entity Resolution .
Take a look to Experiments -> Test_metablocking.java