ˇTransform
each rule into a query
ˇFor
each query:
lSend
to database
lRecord
number of matches
lRetrieve top-k matching documents
ˇAt
the end of round:
lAnalyze matches for each category
lChoose
category to focus on
l
Sampling proceeds in rounds:
In each round, the rules
associated with each node are turned into queries for the database