An introduction to probabilistic graphical models for relational data

We survey some of the recent work on probabilistic graphical models for relational data. The models thatwe describe are all based upon ’graphical models’ [12]. The models can capture statistical correlations
among attributes within a single relational table, between attributes in different tables, and can capture
certain structural properties, such as the expected size of a join between tables. These models can
then be used for a variety of tasks including filling in missing values, data summarization and anomaly
detection. Here we describe two complementary semantics for the models: one approach suited to
making probabilistic statements about individuals and the second approach suited to making statements
about frequencies in relational data. After describing the semantics, we briefly describe algorithms for
automatically constructing the models from an existing (non-probabilistic) database.