Ranking and Classifying Legal Documents using Conceptual Information

A substantial part of all written legal information is published in electronic form these days. Existing retrieval systems, however, are increasingly found to be inadequate. Conceptual ranking and retrieval, in this case based on Bayesian statistics, can be a powerful alternative. Working prototypes of two applications are described. The first one provides the user with the possibility to define, test and save retrieval concepts. Such concepts can be used to rank documents retrieved from a database. The second application reads the saved concepts and calculates the probability that a new document is relevant to the concept.