Storing and analysing fuzzy data from surveys by relational databases and fuzzy logic approaches

Respondents cannot always explicitly state which numeric value or linguistic term is the most suitable to express their opinions. However, the answers are mainly stored in classical relational databases forcing answers to be crisp values. In this way the valuable information related to the vagueness of opinions is lost. In order to avoid this drawback, the data should be stored in a database capable to manage fuzzy data. Adjusting classical relational databases for storing fuzzy data is a promising option which is examined in this paper. Hence, in data analysis the stored fuzziness can be fully exploited. Generally, analyses can be performed to get the summarized information from the data or focused on analyzing particular tuples. Concerning the former, linguistic summaries are extended to cope with fuzzy data. Concerning the later, we are focused on revealing similar entities to the specified one (risqué, optimistic...). Anyway, many other analyses can be performed on stored fuzzy data.