Collecting and managing fuzzy data in statistical relational databases

Abstract

Statistical institutes are focusing on variety of data sources from traditional surveys to big-data. Many of these data and concepts can be expressed as crisp values. But many other data cannot be expressed by precise values. In order to collect, store and manage the fuzziness in data we have adapted the fuzzy meta model as an extension of traditional relational database. Furthermore, experts’ knowledge often contains vagueness and subjectivity. If we store this knowledge in a fuzzy database we can build knowledge management systems capable to cope with fuzziness. Statistical institutes cooperate in the data exchange. We have briefly discussed a simple way of extending the SDMX standard to cope with the fuzzy data in a way that does not influence exchanging precise values. Our research was focused on examining promising ways for managing fuzziness of real world because statistical institutes have been starting to analyze variety of promising data sources where not all data are always