Open problems in relational data clustering

Data clustering is the task of detecting pat-terns in a set of data. Most algorithms
take non-relational data as input and are
sometimes unable to find significant patterns.
Many data sets can include relational infor-
mation, as well as independent object at-
tributes. We believe that clustering with re-
lational data will help find significant pat-
terns where non-relational algorithms fail.
This paper discusses two open problems in
relational data clustering: clustering hetero-
geneous data, and relation selection or ex-
traction. Potential methods for addressing
the problems are presented.