A theoretical model for crisp and fuzzy time

Based on the way time is expressed in natural language, as was studied in the linguistic part of our research, it is observed that time indications usually are rather approximative and situated within the context of the traditional hierarchical time unit system, thereby choosing the most appropriate level of detail. In this paper we present a model for temporal data which aims at being both as flexible and as general as possible. Flexibility is obtained by the use of fuzzy set theory and of granularities which brings the model very close to its natural language counterpart. Generality is achieved by the dynamic aspect of the granularity structure and by the support of spans, time points and intervals, absolute and relative time, aperiodic as well as periodic time elements.