Data dependencies are used in database schema design normalize such a database. It is therefore useful to develop to enforce the correctness of a database as well as to an algorithm that is capable of mining dependencies in reduce redundant data. These dependencies are usually observed data. In this article, we suggest a learning determined from the semantics of the attributes and are method for mining MVDs in observed data. Our system then enforced upon the relations. This article describes requires no a priori knowledge regarding the semantics a bottom-up procedure for discovering multivalued dependencies (MVDs) in observed data without knowing a of the attributes involved. The required input is simply a priori the relationships among the attributes. The prorepository of observed data (preferably in a tabular form).
posed algorithm is an application of the technique we
Our system is capable of learning MVDs from the data designed for learning conditional independencies in and outputs a database schema encoding all the discovprobabilistic reasoning. A prototype system for autoered MVDs. In fact, the output schema satisfies an acyclic mated database schema design has been implemented. Experiments were carried out to demonstrate both the join dependency (Wong, Butz, & Xiang, 1995; Wong, effectiveness and efficiency of our method.