Abstract: A method for extracting a preferred set of textual records from a database includes the following features. Priority values are assigned to each of a plurality of predefined category structures. Textual records are assigned a relevance value with respect to each category structure. If a record's relevance value exceeds a predetermined threshold value, that record is associated with the category structure. Each category has a list of associated textual records which are retrieved. Textual records are selected from the set of retrieved textual records and assembled into a set. Information on how the subscriber uses the set is gathered, and new rankings for the category structure are computed.