Automated adaptation strategies for stream learning

Abstract Automation of machine learning model development is increasingly becoming an established research area. While automated selection and data pre-processing have been studied in depth, there is, however, a gap concerning adaptation strategies when multiple are available. Manually developing strategy can be time consuming costly. In this paper we address issue by proposing the use flexible adaptive mechanism deployment for strategies. Experimental results after using proposed with five algorithms on 36 datasets confirm their viability. These achieve better or comparable performance to custom repeated any single mechanism.