We define the concept of the representative basis for interesting association rules, and an inference system which is purely qualitative. The representative basis is unique, and minimal with respect to the inference system. On the representative basis, the inference system is correct and complete. Experimental results show that the number of rules in the representative basis is significantly reduced with respect to the number of rules generated by other existing approaches.

Using fixpoint operations, we define a new semantics for Datalog programs with negation. The semantics is deterministic. Given a program /spl Delta/, the model of /spl Delta/ in the new semantics includes the well-founded model, and is included in all stable models, if such models exist. In particular, there exist programs that have no stable models, or their well-founded model is empty, but their model in the new semantics is meaningful.

A method for updating deductive databases that allows the insertion or deletion of a fact over intensional predicates in a deterministic manner is presented. It is shown that, contrary to most other approaches, inserting or deleting facts over tensional predicates can always be accomplished without having to make choices. The approach relies on well-founded semantics and on the following two basic approaches: deleted facts are explicitly stored in the database and the inserted and deleted facts may concern any predicate, not just extensional predicates.>

The authors present a novel approach to database updating through universal scheme interfaces. The main contribution of the approach is the elimination of non-determinism. That is, contrary to most other approaches, inserting or deleting a tuple can always be done without having to make any choice at all. Tuples that have been deleted from the database are explicitly stored and are used subsequently in order to invalidate certain derivations. Updates are performed in a monotonous manner, and updates satisfy the property of reversibility.>