Discovery and Maintenance of Functional Dependencies by Independencies

For semantic query optimization one needs detailed knowledgeabout the contents of the database. Traditional techniquesuse static knowledgeabout all possible states of the database whichis already given. Newtechniques use knowledgeonly about the current state of the database whichcan be found by methods of knowledgediscovery in databases. Databases are often very large and permanentlyin use. Therefore, methodsof knowledgediscovery are only allowed to take a small amountof the capacity of the database system. So database access has to be reduced to a minimumduring the process of discovery and obviously if the database changes during the process of maintenanceof the discovered knowledge. In this paper, the maineffort has beenput into minimizing the numberof database accesses, w.r.t, discovery and maintenance. This is exemplified by the discovery of functional dependencies. Weimprovethe inference of functional dependenciesby usingindependencies and cardinality dependencies. First wegive a...

• Type: Conference
• Year: 1995
• Where: KDD
• Authors: Siegfried Bell
• Keywords: Cardinality Dependencies, Data Mining, Database, KDD 1995, Semantic Query Optimization