Use of data analysis techniques for relational query optimization

The problem of relational query optimization requires a broader set of knowledge than generally considered. To acquire knowledge about the use of the database, the authors propose to apply techniques of correspondance and cluster analysis to database queries. First they present the different levels of knowledge that are useful for the query optimization process. Then the authors justify and detail the data analysis approach considered. The interpretation of the output of this analysis provides knowledge about database use. This knowledge is inserted into the knowledge base of an expert system which can then increase its decision capability during a query optimization process.