Towards an integrated methodology for the development of hybrid information systems.
Our modern information society has produced many sophisticated requirements for the
development of information systems. A new challenge is the study of hybrid information
systems that combine traditional information systems with knowledge-based systems. This
new generation information system is considerably more powerful than a simple
extrapolation of existing system concepts. It is easy to imagine the advantages of powerful
knowledge-based systems with efficient access to several large databases, and of large
traditional information systems with added intelligence.
Due to the complex nature of hybrid information systems, it is umealistic to expect that they
can be developed using one standard method. The use of several independently developed
methods has a number of drawbacks, such as inconsistency, redundancy, amount of effort
required and possible loss of information.
In an attempt to provide at least a partial solution to this problem. this thesis describes a new
integrated methodology for developing hybrid information systems. This methodology
combines the method for developing traditional information systems with the method for
developing knowledge-based systems. The new methodology provides a hybrid lifecycle
process model to combine the conventional waterfall process with rapid prototyping and
The proposed methodology integrates four eXlstmg methods using two integration
approaches: intra-process and inter-process. In the requirements analysis phase. a structured
method is applied to function analysis, an information modelling method is applied to data
analysis, and a knowledge acquisition method is applied to knowledge analysis. An intraprocess
approach is then used to integrate these techniques using consistency rules. In the
design phase. the new methodology uses an inter-process approach to transform requirements
analysis to object-oriented design by a transformation algorithm. Finally, an object-oriented
method is applied to the design and implementation of hybrid information systems.
Using the new methodology, a hybrid medical information system for dizziness (HMISD)
was developed, which combines components of traditional medical information systems with
components of medical expert systems. The construction and development of this software
are described in detail. The system can support activities in hospitals including registration,
diagnosis, investigations, drug management and clinical research. It provides assistance to
hospital doctors and general practitioners.
The performance of HMISD is evaluated by testing ninety three real patient cases and taking
two investigations from medical staff and patients. The evaluation results show that HMISD
is of good quality and that most of its users are satisfied.
Three approaches are used to evaluate the proposed methodology: analysis of the
development of HMISD, comparison with existing methodologies using CMD and expert
evaluations. The evaluation conclusions indicate that this new integrated methodology can
take advantage of the four existing methods and also remove some of the limitations of each
individual method. It is applicable to the development of traditional information systems,
knowledge-based systems, and large and complex hybrid information systems.