Pozyskiwanie wiedzy z relacyjnych baz danych: wielopodmiotowe podsumowania lingwistyczne

The aim of this article is to show how fuzzy logic based algorithms can be applied to analyze large datasets and present its results in a human-friendly form: using natural language. A new concept of linguistic summaries is demonstrated: multi-subject linguistic summaries of relational databases, that extends the classic manner. This paper focuses on new, interesting forms of linguistic summaries, which are represented by equations (4), (11), (15) and (18). This article also contains discussion about calculating degrees of truth of the new forms. From the potential end user point of view simplified form of presenting results using natural language expressions is the most important thing. This paper includes demonstration and description of standalone application that generates analysis of large dataset and presents results using short and intuitive expressions in natural language. Possibilities given by the multi-subject linguistic summaries, e.g. description and comparison of more than one subject in one summarization, makes them great extension and complementation of existing forms of linguistic summaries.