In very complex and unstructured domains, the Intelligent Decision Support Systems become very important tools for the expert, since allow to manage a quantity of information in a way that would be impossible to do manually. Inside this kind of systems, the classification tools are one of the most common, and, specifically, the clustering techniques. However, these techniques have problems when managing huge amount of variables and classes, because the interpretation of the generated classes becomes very complicate. For this reason, in this project we want to generate an automatically conceptual interpretation of the classes generated by a clustering technique to help in the labor of the expert with a clearer vision of what is representing each class in order to understand quickly and easy what are the properties and characteristics of these data.