Abstract:

Let C = (E,F,R) a formal context, R is a binary relation defined on ExF  where E is a set of objects and F is a set of attributes. The Galois Lattice (L, <<) generated from R is a good structure and it is used in several areas of Computer Sciences such as Object Oriented Database, Knowledge Data Discovery, Learning etc. The number of Galois Lattice Concepts is exponential 2^|F|, where |F| is the cardinality of attributes. Some research works shows limits of the use of this structure when the number of attributes increases. For example, in Data mining when the number of transactions is very large, in Information Retrieval when we have to represent a very large documentary database. In this paper, we present a method which allows us to generate a subset of concepts from the set of Galois Lattice Concepts and in the same time we ensure that the data represented by Galois Lattice is covered. We apply this method to define an Intermediate Representation (IR) of a documentary database in the objective to give a synthetic view of its content for the user. This IR will be used by the user during a retrieval session. Thus, we visualize this IR as a graph and the user could navigate in it and expresses their queries by exploiting the semantic of this graph.