Fuzzy database modeling (Q1277799)

Many applications characteristically require the modeling of complex data types, which has led to new database models other than the established relational paradigm. If applications use imprecise or fuzzy information or vague quantifiers like 'many', 'some' etc. data models must comprise fuzzy data. The book is subdivided into an introduction and four chapters, each of them coming with its own abstract, keyword list and bibliography. The introduction gives a short overview of fuzzy sets, similarity relations and the three levels of uncertainty -- enterprise, data and query language. Chapter 2 is concerned with the physical design of fuzzy databases, that is, how to realize efficient access to precise and fuzzy data at the same time. The Multi Level Grid File data structure (MLGF) as a generalization of Grid Files is introduced and compared with other proposed structures, the results of performance tests are shown. Perhaps the reader should now proceed with chapter 4 on logical database models for uncertain data. This is because the Extended NF\(^2\) Model and the Fuzzy Object Oriented Model (FOOD) discussed here are the targets of transforming conceptual schemas to logical ones in chapter 3. For Extended NF\(^2\) a query language and query processing algorithms are presented. FOOD is exemplified by a an explicit application. Further sections discuss fuzzy object algebra, coupling with fuzzy logic and a formal description of fuzzy deductive object oriented database models. Chapter 3 is on conceptual modeling and extends the IFO Data Model in two ways corresponding to uncertainty at the attribute level (ExIFO) and at the class and entity level (ExIFO\(_2\)). At the attribute level the possibility of incomplete-valued attributes like 'salary between 5000 and 6000' or fuzzy-valued attributes -- 'colour is something like blue or green' -- gives rise to two new constructors of entity types. The authors give rules for constructing correct ExIFO schemas and for their transformation into Extended NF\(^2\). Fuzziness in connection with object orientation is the object of ExIFO\(_2\), which is transformed into FOOD. The last chapter investigates fuzzy relational databases and focusses on fuzzy attributes, fuzzy constraints and fuzzy functional or multivalued dependencies, which are defined by conformance of tupel. Several implementations based on Prolog, Visual Basic, C++, POPLOG e.a. are presented with details of code. Nearly every new concept is clarified with well chosen examples, making it easily accessible to graduate students. At the same time the book is giving up-to-date and state of the art information to the database researcher and the IT professional.