Research Of Association Rules Algorithm And Knowledge Management Model Based On Ontology

With the development and popularization of Internet, people need to face a flood of information resources, information content and the level of diversity and complexity, especially when people want to gain a small amount of needed and useful information from much information, they sometimes feel helpless when face such question. So data mining technology research and knowledge management is the objective requirements of the development of global information.In this paper, mainly basing on ontology and making use of ontology's hierarchy, and an algorithm ML_AR which ontology-based multi-level association rules is proposed. The smallest degree of support which is the most critical factor in associate rules, which is used to reduce the search space and limit the number of generated rules, if only a single minimum support, it will implicitly assume that the sub-items in the database have the same a similar nature or frequency of occurrence, in order to address the above issues, and a formula is given which accords to the upper support to calculate the lower level support. Because of the support is decreasing which may cause the parent layer of non-frequent item entry and sub-layer may be frequent, so can regain the items than meet these conditions. If the support of calculated is greater than the of support parent, and the indicating support for selecting about the top is too small which does not meet the actual data, and it should re-adjust the support from the initial point of view. This algorithm has also certain evaluation of the choice of the initial degree of support. Another, as for the problem of giant of unrelated information that emerges when searching knowledge on the knowledge management system, the paper proposes a new knowledge management system framework. The framework makes use of the characteristics of concept of norms, rich semantic and level relations of the ontology and combines the knowledge warehouse of ontology and the first order predicate reasoning engine on the user's query to process the result and make users get ultimately information that matches with search requests in a higher state. Experimental results show that the algorithm is efficient on the time complexity and the space complexity, but also has good performance.