Randomization and entropy in data processing, dynamic systems, machine learning. Popkov Yu.S.

This book is based on randomization and entropy. The combination of these two approaches that arose at the beginning of the century turned out to be very fruitful for solving various problems. This book contains the results of studies of the last ten years, formulating the general concept of entropy-degraded assessment and its use in the tasks of reduction of data dimension, “hard” and “soft” randomized machine learning, forecasting and clustering. Significant attention is paid to the relatively new section of the theory of dynamic (controlled) systems with an entropy operator. Many theoretical results are brought to computational schemes and algorithms. Undoubted interest for the reader can be represented by the theory of entropy-degraded assessment and forecasting of the dynamics of the world population, the daily electric load of the energy system and the evolution of the area of ​​thermocarst lakes.

The book can be useful for students, graduate students and scientists interested in the concept entropy-degree assessment, as well as its practical applications.