Representation And Analysis Of Ontology Changes

As a model which can describe concept hierarchy and semantics efficiently, ontology is widely used in many domain of computer science. However, business dynamics and changes in the environment require that ontology should be changed, so the problem of ontology evolution comes. This is especially true for Semantic Web applications, which are based on heterogeneous and highly distributed information resources.Ontology evolution is very important in ontology development. Ontologies often require be changed, however, as the size and the number of ontologies increase, the costs to adapt them to the changing requirements increase more. Developing ontologies and their applications is expensive, but evolving them is even more expensive. The experiences show that the traditional software system's maintenance costs exceed the development costs by a factor of between two and four. This is no different for ontologies. The costs can be even higher, due to the collaborative development of ontologies and their physical distribution.Ontology evolution is the timely adaptation of an ontology to the arisen changes and the consistent propagation of these changes to dependent artefacts. Since a change in the ontology can cause inconsistencies in other parts of the ontology, as well as in the dependent artefacts, the ontology evolution has to be considered as a process and it is driven by changes. Ontology evolution process must realize two crucial functions: (1) handling the required ontology changes; (2) ensuring the consistency of the underlying ontology and all dependent artefacts. In addition, it should support the user to manage changes more easily and offer advice to the user for continual ontology refinement.The ontology evolution process which is accepted abroad includes six phase: (1) change capturing; (2) change representation; (3) semantics of change; (4) change propagation; (5) change implementation; (6) change validation. The core ontology evolution process realizes the two crucial functions through the four phases from (2) to (5).Our primary intention is to analyze the methods of ontology change representation and the effects of corresponding change operations, and give a form for representing ontology changes which can represent more clearly and improve the ontology evolution process. Change representation is the first phase of the ontology evolution process, and it is the basis for handling changes. To resolve changes, they have to be identified and represented in suitable format. There are many patterns of ontology changes, and each may bring different effects. To a change requirement, the more clearly it is represented, the more correctly and efficiently it can be resolved.Another aspect here is to analyze the realization of the core ontology evolution process. We will mainly work on the evolution within an ontology. In addition to the change representation phase, that will refer to the phases of semantics of change and change implementation.The task of semantics of change phase is to enable resolution of induced changes in systematic manner, ensuring consistency of the whole ontology. This phase realizes the core function of resolving changes within one ontology. In this phase, we discuss the ontology evolution strategy firstly, and then, analyze another method based on it: considering the ontology evolution as a graph search process on the ontology consistency model. We will (1) define the ontology consistency model and consistency constraints, (2) analyze change operations on the situation of constraint violations and the corresponding treatment strategy and (3) discuss the efficiency of this graph search process.The role of the change implementation phase is (1) to inform an ontology engineer about all consequences of a change request, (2) to apply all the (required and derived) changes and (3) to keep track about performed changes. Here, we mainly discuss (3). To resolve the problem of change propagation, it is often needed to know the history of changes, and change logging is a main method to get the history information. Obviously, change logging has great relation to change representation. If we want to get enough change information, we should represent changes clearly.Through the discussion of ontology evolution process, we analyze the effects and resolution strategies of our change representation, showing that this change representation can bring many advantages, for example, it can reduce redundant change operations and improve the veracity of evolution process.According to our analysis, we can see that for the semantic model of ontology, changes which represent depending on semantics may be easier to understand to people and better to deal with the computer. This can improve the effect of the ontology evolution process to a great extent.