No label defined (Q4219293)

The book is an attempt to cover a broad range of diverse database and knowledge-based systems. The idea to integrate both these systems is not now despite of the fact that both areas have been developed relatively independently in the past. Recent approaches to knowledge-based systems require a database techniques to increasing their practical usage, on the other hand, today's database systems extend their power absorbing various techniques originated in knowledge representation research. The book is divided into five parts. The first part summarizes basic notions and results concerning conceptual modelling, relational databases, and object databases. In Chapter 1 the author presents shortly E-R modelling and agent-object-relationship modelling as easy tools useful for formalizing an application domain. Chapter 2 introduces the relational data model as the most fundamental case of knowledge systems. Relational queries, integrity constraints, and updates are presented with the help of logic, including such problems as null values and Reiter's completion theory. Chapter 3 concerns today's alternative to relational systems called object-relational approach. The addition of abstract data types (ADT) to relations makes it possible to define ADT-databases with rich data structures going beyond flat first normal relations. Part 5 of the book extends a database system by rules. Chapter 4 and 5 introduce so-called reaction and deduction rules. The reaction rules subsume production rules, triggers, and commitment rules that are well-known from expert systems, database systems, and agent-oriented programming, respectively. Deduction rules are explained in the environment of DATALOG. Besides its basic version, extensions such stratifiable, well-founded, and stable model are defined. Part 3 provides four important chapters about positive knowledge systems. In these systems information units are always consistent with each other. In addition to previous systems, the author discusses temporal and fuzzy databases (Chapter 7 and 8). Besides basic models, associated operations and deduction rules are shown in both cases. Chapter 9 offers further examples of positive knowledge systems: multi-level secure, lineage, disjunctive and S5-epistemic databases. Part 4 deals with admitting negative information. In Chapter 10, the author shortly introduces principles of non-positive knowledge systems. Nice examples of this class of systems follow. Chapter 11 defines so called relational bitables and relational factbase. In bitables, positive and negative information is stored in different kinds of tables. Possibilistic databases described in Chapter 12 extend fuzzy databases in the same way as relational factbases extend relational databases. Further examples are contained in Chapter 13. Final part 5 gives more information on reaction and deduction rules. Chapter 14 defines the operational semantics of reaction rules and shows how distributed query answering and updating in multidatabases and co-operative query answering protocols can be formally specified and verified. In Chapter 15, elements of general closure semantics for deduction rules in knowledge bases are developed. The key concept of a deductive knowledge base is given in a new and direct approach. Finally, in Chapter 16, some new services of advanced knowledge systems are mentioned. In addition to query answering, these features include active input processing by means reaction rules, abductive generation of explanations, diagnoses, configurations, and plans. Three appendixes summarizing selected theoretical notions conclude the book. The book is written on a sufficient formal level in such way that a reader, not too much experienced in various logical approaches, can find it interesting and beneficial. Also some new models of existing notions bring a contribution. I can recommend the book to all people from database and knowledge communities who want to move their research and practice to more advanced innovative systems.