Generation of interval-valued fuzzy sets in machine learning

In machine learning, the aim is to build a model from a training set. However, such training sets rarely cover the entire description space of the possible cases of the considered problem. Such a lack of knowledge leads to imperfect models and errors during its use (wrong classifications or predictions). To overcome such a lack of knowledge, an approach is introduced to construct intervalvalued fuzzy sets in a preparation step of the training set.