4. Conclusion

Different approaches to building conceptual classifications from Web corpora have been presented. To perform such a task, it is necessary to apply a comprehensive text mining process. In this process, two essential steps were highlighted. The first phase of the overall text mining process described concerns the acquisition of the corpus from the Web. Such an approach is not always easy, especially when the aim is to build a specific conceptual classification. This requires the acquisition of specialized corpora. After this initial phase, text standardization and terminology extraction steps (as described in ) must be applied. The next phase consists of implementing methods and algorithms to propose conceptual classes to the user. We have presented a precise description of mixed methods (statistical and linguistic) as well as purely statistical approaches for the construction...