Triadic FCA (3FCA) is a natural extension of FCA designed to handle three dimensional data sets. This approach has been later extended to n-adic data sets. One major problem, especially for larger data sets is to navigate among the knowledge patterns as are they revealed by computing concepts. This can be solved, in 3FCA, by a local approach, by projecting along a dimension and navigate in the “flattened” concept lattices which are arising by this procedure. Another challenge is to find a specific subset of concepts based on some user search criteria, which in fact are encoded in a set of constraints. These constraints are defining a knowledge base on which a Logical Programming tool, Answer Set Programming (ASP), is used to compute quickly the answer set.
On the other hand, we can apply all these results to investigate user behavior by mining “knowledge gems” from weblogs. By this approach, one can highlight the conceptual structure of educational resources (“how does the educator structure its own e-learning resources”) and then proceed in a stepwise unfolding of several structures: attractors – structures to which users adhere and which are considerably influencing their browsing behavior, user life tracks, analysis of bundle of users, detecting pioneers, etc.