A Relation Combination Model for Knowledge Maps

Knowledge maps are useful visualization tools in the field of knowledge management. They are typically used to help users locating knowledge or information by providing relations in various forms. Different stakeholders may have different interests (requirements) while using the knowledge or information stored in a source. Moreover, the interests may change over time. However, current researches on knowledge maps are not sufficient for complicated and ever-changing requirements. Some knowledge maps merely demonstrate the relations previously stored in the knowledge bases or databases, which are limited when more abstract semantics are need to be shown; Others build up their special relation structures for specific purposes, but usually result in duplicate manual efforts and redundant in the storage. In this paper a relation combination model is proposed, using a set of relation operators (cascading operator and logic operators) to combine or reorganize the existing relations into more abstract ones. Through re-organization or combination, the relations can be visualized in different views or different abstractions, which are more useful to different stakeholders with different interests when they use the same set of information. The results will be quickly obtained without any coding and re-input of data. Thus the manual effort is reduced and the possible inconsistency between data sources is avoided. The model and an interactive user interface have been implemented upon a knowledge management platform. Finally, a case study is given and discussed.