Analysis of methods for objects selection under specified conditions

The problem of clustering objects under several conditions is frequently presented. The selection has been made in the past using statistical methods for discriminating certain parameters or creating queries from a database, which looks more practical. In general queries are created using the SQL method. The classical SQL methodology using crisp qualifiers causes difficulties in some decision making processes especially when it is mandatory to move to the definition of practical indicators or categories and to evaluate them according to certain practical assumptions. Recently fuzzy logic has been embedded in SQL to improve its performance, but the applications have been basically oriented to analysis of word similarity and indexing. The paper analyzes the statistical, SQL, and fuzzy SQL methods, presenting a practical application classifying several companies using the last one.