Semantic wikis enable collaboration between human agents and the creation of associated knowledge systems. In this way, data embedded in semantic wikis can be mined and the resulting knowledge patterns can be reused to extend and improve the structure the wiki contents. This paper proposes a method for guiding the reengineering and improving the structure of semantic wikis. This method suggests the creation of categories and relations between categories using Formal Concept Analysis (FCA) and the Relational Concept Analysis (RCA) extension. FCA allows the design of a concept lattice while RCA provides relational attributes completing the content of formal concepts. The originality of the approach is to consider the wiki content from FCA and RCA points of view and then to extract knowledge units from this content allowing a factorization and a reengineering of the wiki structure. This method is general and does not depend on any domain and can be thus generalized to every kind of semantic wiki. Examples are studied throughout the paper and experiments show the promising results.