Inference aggregation detection in database management systems

Abstract

The author identifies inference aggregation and cardinality aggregation as two distinct aspects of the aggregation problem. He develops the concept of a semantic relationship graph to describe the relationships between data and then presents inference aggregation as the problem of finding alternative paths between vertices on the graph. He presents an algorithm for processing the semantic relationship graph to discover whether potential inference aggregation problems exist. A method of detecting some aggregation conditions within the database management system (DBMS) is presented that uses the normal DBMS query language and adds additional catalytic data to the DBMS to permit a query to make the inference. The author also suggests the use of set theory to describe aggregation conditions and the addition of set operations to the DBMS to permit the description of aggregation detection queries.<>