Autonomous learning from examples

Machine learning is one of the major challenges of Artificial Intelligence. Inductive reasoning on the basis of classified examples is at the origin of many data mining methods, so fashionable today in the area of massive data processing. The function of such methods is to intensively describe a collection of concepts, initially expressed in extensional form by means of a set of classified examples. Most methods have as their starting point a fixed collection of examples, which is considered sufficiently meaningful. However, in the vast majority of relatively complex problems, it will not always be possible to have such an initial collection. Moreover, many examples may still be undiscovered and some of them are incompletely specified. This book proposes a solution for such situations, based on an autonomous learning model guided by an incremental induction-deductive reasoning process and by the experiences of the model itself and the user, with no limits on the number of examples to be explored.