Reinforcement Learning, Unsupervised Methods, and Concept Drift in Stream Learning

András A. Benczúr, Levente Kocsis, Róbert Pálovics

Abstract: In this chapter, we give a brief overview of a few special topics in online machine learning, all of which are extensively covered in recent surveys. In Section “Reinforcement Learning,” we survey reinforcement learning. In Section “Unsupervised Data Mining,” we describe unsupervised data mining methods, including clustering, frequent itemset mining, dimensionality reduction, and topic modeling. In Section “Concept Drift and Adaptive Learning,” we describe the notion of the dataset drift or, in other terms, concept drift and list the most important drift adapting methods. We only discuss representative results in these areas. This chapter is an extension of the other chapters in this handbook, “Overview of Online Machine Learning in Big Data Streams,” “Online Machine Learning Algorithms Over Data Streams,” and “Recommender Systems Over Data Streams.”