| Concept lattice is a kind of conceptual hierarchical structure based on the object and attribute relation in data set. It expresses the general and special relation between concepts vividly. Concept lattice was presented by German mathematician Will.R and has already been used in data mining, information retrieval and software engineering. The main work of this paper is the theory of concept lattice and its application in data mining. We want to explore something in this area, on one behalf, dig the potential ability of concept lattice's application, on the other behalf, and find some new data mining algorithms.The innovation of this article:1,Improve the attribute reduction in concept lattice theory since it eliminates the redundant computation maximally and speed the reduction efficiency.2,We discuss the building of concept lattice which is core topic in the theory of concept lattice and put forward a building method based on attribute matrix. This algorithm considers the same objects as a whole and improves the building efficiency. At the same time, the process of building is also the process of attribute reduction, thus avoid the redundant process between building and reduction.3,Apply the concept lattice to extract rules including classification and association rules. The algorithm is based on the idea of concept lattice, but it avoid the building time consuming and can be consider as a quick time algorithm. The result of the experiment we use the rules to classify is desirable.4,We use concept lattice to cluster documents which is one of the innovations of the paper. As a natural tool to cluster, concept lattice still has some big problems to tackle; we put forward these peanuts and crack some of them. Lastly, we transfer the image data to formal context which the rule extraction algorithm we discussed can be used.At last, this paper concludes the basic theory of concept lattice and its application in data mining, at the same time, prospect new directions about the future work. |
PDF Full Text Request