Benchmarking concept drift adoption strategies for high speed data stream mining

Abstract

Data streams are significantly influenced by the notion change that is termed as concept drift. The act of knowledge discovery from the data streams under notion adaption is a significant act to achieve the conventional learning of the streaming data. The concept drift for conventional learning of streaming data can be done under set of notions that can be either static or dynamic. Due to the large scope of concept drift that spanned to different domain contexts of data streaming, the existing models are partially or fully not generalized and compatible to different streaming and notion change context. In this context, this paper presents the review of these models that includes nomenclature of mining streaming data and notion evolution in concept drift adoption strategies.