Extracting Knowledge Context Patterns from Instances

Knowledge management is recognized as an important means for organization to gain and sustain competitive advantage. One of objectives of knowledge management system (KMS) is to make the knowledge available to the right worker at the right time and right place. To this end, knowledge context is crucial. The key issue in current research is how to facilitate the use of contextual information in KMS. This paper focuses on extracting knowledge context patterns from instances. Context pattern is description of a context class which characterizes the situation where one knowledge item or one class of knowledge items is needed. We argue that context pattern should be general enough to reflect common characteristic of the context class and unique enough to distinguish the corresponding knowledge item (or knowledge type) from others. This paper presents a method to mine the patterns from instances. Two experiments are conducted to evaluate the effectiveness of the approach. The result shows that the method is effective in appropriate environment.