Enhancing Concept Drift Detection with Simulated Recurrence
This paper focuses on the concept drift detection and proposes how to extend the functionality of a statistical concept drift detector for unlabeled observations. For those algorithms the previously developed approach so-called simulated recurrence is implemented as a separate module. It provides information regarding the possible data distribution after concept drift detection. The proposed approaches were compared with five detection algorithms on the basis of computer experiments which were carried ut on the UCI benchmark datasets.
KeywordsDetection Module Concept Drift Reference Dataset Data Window Multivariate Version
Unable to display preview. Download preview PDF.
- 1.Newman, D.J., Asuncion, A.: UCI machine learning repository (2007)Google Scholar
- 3.Friedman, J., Rafsky, L.: Multivariate generalizations of the wald-wolfowitz and smirnov two-sample tests. The Annals of Statistics, 697–717 (1979)Google Scholar
- 7.Hulten, G., Spencer, L., Domingos, P.: Mining time-changing data streams. In: ACM SIGKDD Intl. Conf. on Knowledge Discovery and Data Mining, pp. 97–106 (2001)Google Scholar
- 9.Sheskin, D.J.: Handbook of Parametric and Nonparametric Statistical Procedures, 4th edn. Chapman & Hall/CRC (2007)Google Scholar
- 12.Vreeken, J., van Leeuwen, M., Siebes, A.: Characterising the difference. In: Berkhin, P., Caruana, R., Wu, X. (eds.) KDD, pp. 765–774. ACM (2007)Google Scholar