Improving the learning of recurring concepts through high-level fuzzy contexts

In data stream classification the problem of recurring concepts is a special case of concept drift where the underlying concepts may reappear. Several methods have been proposed to learn in the presence of concept drift, but few consider recurring concepts and context integration. To address these issues, we presented a method that stores previously learned models along with context information of that learning period. When concepts recur, the appropriate model is reused, avoiding relearning a previously seen concept. In this work, in order to model the vagueness and uncertainty associated with context, we propose the inference of high-level fuzzy contexts from fuzzy logic rules, where the conditions result from fuzzified context inputs. We also present the changes required for our method to deal with this new representation, extending the approach to handle uncertain contexts.