- 726 Downloads
This introductory chapter intends to present the challenges related to the problem of learning from data streams in nonstationary environments. It focuses on the major challenges related to the learning with concept drift, learning with concept evolution, and learning with both concept drift and concept evolution. Then, it classifies the different methods and techniques of the state of the art that are used to address these challenges. This categorization is achieved according to how the learning is performed, how the data streams are processed, and how the changes are detected and integrated into the model. Finally, this chapter ends with a compact summary of the contents of this book by providing a paragraph about each of the contributions and how the learning process from data streams is performed (single learner or ensemble learners, centralized processing or distributed computing, classification, regression or clustering, window-based or sequential-based, applications targeted, etc.).
KeywordsConcept Drift Nonstationary Environments (NSE) Data Stream Clustering Method Dynamic Weighted Majority (DWM) ADWIN Bagging
- 5.Mouchaweh, M.S.: Diagnosis in real time for evolutionary processes in using pattern recognition and possibility theory. Int. J. Comput. Cogn. 2(1), 79–112 (2004)Google Scholar
- 7.Hulten, G., Spencer, L., Domingos, P.: Mining time-changing data streams. In: Proceedings of the Seventh ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 97–106. ACM, New York (2001)Google Scholar