This monograph is devoted to the consideration of new deep neural networks – Evolving Hybrid Stacking Neuro-Neo-Fuzzy Systems of Artificial Intelligence based on Group Method of Data Handling, which, in turn, is the first known method of deep learning. This method is based on the principle of self-organization and, unlike in other deep learning methods, it allows not only to adjust the weights of neural connections, but also to synthesize the optimal network structure in online mode. The proposed approach allows you to process information in online mode and solve a wide class of Data Stream Mining problems. Particular attention in the book is paid to the online bagging approach, when optimal accuracy results are synthesized for solving the problems of pattern recognition, forecasting, and classification.
The book is aimed primarily at specialists in the field of deep learning involved in the development of new architectures and algorithms for deep learning networks and their application in forecasting, pattern recognition and medical diagnostics, but will also be of use to students of computer science and AI, and the general scientific community at large.
