### abstract ###
Many databases store data in relational format, with different types of entities and information about links between the entities
The field of statistical-relational learning (SRL) has developed a number of new statistical models for such data
In this paper we focus on learning class-level or first-order dependencies, which model the general database statistics over attributes of linked objects and links (e g , the percentage of A grades given in computer science classes)
Class-level statistical relationships are important in themselves, and they support applications like policy making, strategic planning, and query optimization
Most current SRL methods find class-level dependencies, but their main task is to support instance-level predictions about the attributes or links of specific entities
We focus only on class-level prediction, and describe algorithms for learning class-level models that are orders of magnitude faster for this task
Our algorithms learn Bayes nets with relational structure, leveraging the efficiency of single-table nonrelational Bayes net learners
An evaluation of our methods on three data sets shows that they are computationally feasible for realistic table sizes, and that the learned structures represent the statistical information in the databases well
After learning compiles the database statistics into a Bayes net, querying these statistics via Bayes net inference is faster than with SQL queries, and does not depend on the size of the database
### introduction ###
Many real-world applications store data in relational format, with different tables for entities and their links
Standard machine learning techniques are applied to data stored in a single table, that is, in nonrelational, propositional or ``flat" format  CITATION
The field of statistical-relational learning (SRL) aims to extend machine learning algorithms to relational data  CITATION
One of the major machine learning tasks is to use data to build a  generative statistical model  for the variables in an application domain  CITATION
In the single-table learning setting, the goal is often to represent predictive dependencies between the attributes of a single individual (e g , between the intelligence and ranking of a student)
In the SRL setting, the goal is often to represent, in addition, dependencies between attributes of different individuals that are related or linked to each other (e g , between the intelligence of a student and the difficulty of a course given that the student is registered in the course)
