Research On Temporal Data Model With Time Indeterminacy

Information system is the foundation of the knowledge economy and main feature of future societal economy. The studies of the medical information system (MIS) will accelerate the advancement of technologies and the reform of our society. The temporal information that includes natural language expressions is the key to MIS. However, it is not fully solved on how to process the indeterminate temporal information in the natural languages.According to the need of MIS, the current study technologies of temporal data models and some classical indeterminate temporal medical information data models are studied in this paper. Through the analyses of the models such as GCH-OODM model, this paper proposes an indeterminate temporal data model, named NLTM, based on natural language expressions, to deal with the above problems in the medical information system.First of all, all sorts of temporal primitive types are extended to support indeterminacy in this paper. The representations of indeterminate temporal primitive types are redefined, union, overlap, and difference operators are extended, set operation and relational calculus are given between indeterminate time point and time interval, and formalization definition of several temporal relation algebra operations are given.Secondly, based on natural language expressions, a temporal data model maintaining indeterminate time is proposed in this paper. Time point, time interval, duration and temporal element used in the model are presented, respectively. The model contains three attributes, which are patient ID, patient case history and corresponding time information. The time attribute is divided into date and time-of-date.Thirdly, according to different temporal expressions, all sorts of representations of time attribute are denoted. Depending on the attributedefinitions of NLTM model, the syntax, semantics, and corresponding predicates, functions are described, in which predicates, functions are utilized to support different query manners of NLTM model.Finally, this paper adopts GCH-OODM model and NLTM model to represent the same medical data cases and compare the difference of the two models. Through the rational analyses, the storage performance of NLTM model is better than GCH-OODM model when dealing with indeterminate temporal data. Besides, this paper introduces how to design NLTM model using object-oriented methods.