The article presents a method for improving classification of streaming data influenced by concept shift. For this purpose the algorithms designed for recurring concept drift environments are adapted. To minimize classification error after concept shift, an artificial recurrence is implemented serving as a better starting point for classification. Three popular algorithms are tested on three different scenarios and their performance is compared with and without the application of an artificial recurrence. © 2011 Springer-Verlag.
CITATION STYLE
Sobolewski, P., & Woźniak, M. (2011). Artificial recurrence for classification of streaming data with concept shift. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6943 LNAI, pp. 76–87). https://doi.org/10.1007/978-3-642-23857-4_11
Mendeley helps you to discover research relevant for your work.