Introduction and Overview of the Main Results of the Book

Abstract

In recent decades we are observing an exponential increase in the available digital data, generated in various areas of human activity. This growth is much faster with respect to the increase in the available processing capabilities. Apart from large volumes, the data produced by modern data sources are often dynamic and generated at very high rates. Therefore, there is a big challenge to design new data mining algorithms able to deal with such a streaming nature of data. Data stream mining became a very important domain of computer science and finds applications in many areas, e.g. in engineering and industrial processes, robotics, sensor networks, social networks, spam filtering or credit card transaction flows. In this book we present a unique approach to data stream mining problems, putting emphasis on the theoretical backgrounds of considered algorithms. Contrary to the vast majority of the previously presented in the literature heuristic methods, this book focuses on algorithms which are mathematically justified. However, it should be noted that the heuristic solutions cannot be completely abandoned since they often lead to satisfactory practical results. Therefore, the mathematically justified algorithms presented in this book are sometimes slightly tuned and modified in a heuristic way to increase their final accuracy.