Paired Evaluators Method to Track Concept Drift: An Application in Finance
We consider the problem of forecasting under the environments of sudden unexpected changes. The objective of the forecasting is to detect several different types of changes and to be adaptive to these changes in the automated way. The main contribution of this paper is a development of a novel forecast method based on paired evaluators, the stable evaluator and the reactive evaluator, that are good at dealing with consecutive concept drifts. A potential application of such drifts is Finance. Our back-testing using financial data in US demonstrates that our forecasting method is effective and robust against several sudden changes in financial markets including the late-2000s recessions.
Unable to display preview. Download preview PDF.
- 1.Bach, S., Maloof, M.: Paired learners for concept drift. In: Proceedings of the 8th IEEE International Conference on Data Mining, pp. 23–32 (2008)Google Scholar
- 2.Bifet, A., Gavalda, R.: Learning from time-changing data with adaptive windowing. In: Proceedings of SIAM International Conference on Data Mining (SDM 20007), pp. 443–448 (2007)Google Scholar
- 3.Bifet, A., Holmes, G., Pfahringer, B., Kirkby, R., Gavalda, R.: New ensemble methods for evolving data streams. In: Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 139–148 (2009)Google Scholar
- 7.French, K.R.: Fama/french factors in u.s. research returns data, http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html (accessed July 23, 2010)
- 11.Kuncheva, L.I., Žliobaitė, I.: On the window size for classification in changing environments. Intelligent Data Analysis 13(6), 861–872 (2009)Google Scholar
- 12.Lazarescu, M.M., Venkatesh, S., Bui, H.H.: Using multiple windows to track concept drift. Intelligent Data Analysis 8(1), 29–59 (2004)Google Scholar
- 15.Nara, Y., Ohsawa, Y.: Tools for shifting human context into disasters: a case-based guideline for computer-aided earthquake proofs. In: Proceedings of the 4th International Conference on Knowledge-Based Intelligent Engineering Systems and Allied Technologies, pp. 655–658 (2000)Google Scholar
- 16.Nisida, K., Yamauchi, K.: Learning and detecting concept drift with two online classifiers. In: Proceedings of the 22nd Annual Conference of the Japanese Society for Artificial Intelligence, pp. 3C2–1 (2008)Google Scholar
- 19.So, J., Furuhata, M., Mizuta, T.: Operational model considering transaction costs to correspond to sudden changes in japanese stock markets. In: Proceedings of the 6th workshop SIG-FIN in the Japanese Society for Artificial Intelligence, pp. 23–29 (2010)Google Scholar
- 20.Žliobaitė, I.: Learning under concept drift: an overview. Technical report, Vilnius University, Faculty of Mathematics and Informatics (2009)Google Scholar
- 21.Widmer, G., Kubat, M.: Learning in the presence of concept drift and hidden contexts. Machine Learning 23(1), 69–101 (1996)Google Scholar