Data summarization using extended concept hierarchies

The paper introduces an original approach to data summarization, using concept hierarchies and fuzzy techniques. Only the summarization of precise data obtained from a classical relational database is considered. The proposed method aims at producing summaries which are directly understandable, and therefore described by means of words, or linguistic labels, taken from extended concept hierarchies built on each attribute. The summaries are produced at different levels of granularity, according to the concept hierarchies. This kind of summary can be useful to answer flexible queries, when only a global view of the information stored in the database is needed. Indeed, querying the base of summaries, with the generalization level used in the query, is faster than querying the whole database, and it gives an answer using the same level of vocabulary, i.e., the same imprecision as the query