An experiment in discovering association rules in the legal domain

The article explores the applicability of an algorithm designed for finding association rules in large databases to the discovery of relevant associations from a large case base. More specifically, it describes an experiment intended to explore the applicability of data mining techniques to legal databases. In many areas of law, especially administrative law, many thousands of cases are decided. This data presents a significant resource. We would generally wish to assume that some rule is being followed so that like cases are decided in a like manner. Is there a way of deciding what the rule being followed is from an automated consideration of the data?