Journal Article10.1109/MCI.2015.2471196
Learning in Nonstationary Environments: A Survey
TL;DR: In such nonstationary environments, where the probabilistic properties of the data change over time, a non-adaptive model trained under the false stationarity assumption is bound to become obsolete in time, and perform sub-optimally at best, or fail catastrophically at worst.
read more
Abstract: The prevalence of mobile phones, the internet-of-things technology, and networks of sensors has led to an enormous and ever increasing amount of data that are now more commonly available in a streaming fashion [1]-[5]. Often, it is assumed - either implicitly or explicitly - that the process generating such a stream of data is stationary, that is, the data are drawn from a fixed, albeit unknown probability distribution. In many real-world scenarios, however, such an assumption is simply not true, and the underlying process generating the data stream is characterized by an intrinsic nonstationary (or evolving or drifting) phenomenon. The nonstationarity can be due, for example, to seasonality or periodicity effects, changes in the users' habits or preferences, hardware or software faults affecting a cyber-physical system, thermal drifts or aging effects in sensors. In such nonstationary environments, where the probabilistic properties of the data change over time, a non-adaptive model trained under the false stationarity assumption is bound to become obsolete in time, and perform sub-optimally at best, or fail catastrophically at worst.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
Continual lifelong learning with neural networks: A review.
TL;DR: This review critically summarize the main challenges linked to lifelong learning for artificial learning systems and compare existing neural network approaches that alleviate, to different extents, catastrophic forgetting.
3.2K
Ensemble learning for data stream analysis
TL;DR: This paper surveys research on ensembles for data stream classification as well as regression tasks and discusses advanced learning concepts such as imbalanced data streams, novelty detection, active and semi-supervised learning, complex data representations and structured outputs.
1K
Learning under Concept Drift: A Review
TL;DR: A high quality, instructive review of current research developments and trends in the concept drift field is conducted, and a framework of learning under concept drift is established including three main components: concept drift detection, concept drift understanding, and concept drift adaptation.
995
Learning under Concept Drift: A Review
TL;DR: In this paper, the authors present a review of the recent research in the field of concept drift and propose a framework of learning under concept drift. But, the focus of this survey is on the detection, understanding and adaptation of the concept drift in streaming data.
752
A Survey on Ensemble Learning for Data Stream Classification
TL;DR: This work proposes a taxonomy for data stream ensemble learning as derived from reviewing over 60 algorithms, and important aspects such as combination, diversity, and dynamic updates, are thoroughly discussed.
597
References
Early Drift Detection Method
Manuel Baena-Garc,Jose del Campo ¶ Avila,Albert Bifet,Ricard Gavald,Rafael Morales-Bueno +4 more
- 01 Jan 2005
TL;DR: A method for detecting concept drift, even in the case of slow gradual change, based on the estimated distribution of the distances between classiflcation errors that can be used with any learning algorithm in two ways: using it as a wrapper of a batch learning algorithm or implementing it inside an incremental and online algorithm.
A Regularized Multiple Criteria Linear Program for Classification
Yong Shi,Yingjie Tian,Xiaojun Chen,Peng Zhang +3 more
- 28 Oct 2007
TL;DR: The paper first adds some regularization terms in the objective function of the known multiple criteria linear program (MCLP) model for possible existence of solution, and describes the mathematical framework of the solvability.
The algorithm APT to classify in concurrence of latency and drift
Georg Krempl
- 29 Oct 2011
TL;DR: An explicit drift model is contributed, which is characterising a mixture of independently evolving sub-populations drifting over time, and an arbitrary sub-population tracker algorithm is presented, which can track and predict the distributions by the use of unlabelled data.
Classifier Ensembles for Changing Environments
Ludmila I. Kuncheva
- 09 Jun 2004
TL;DR: This work considers strategies for building classifier ensembles for non-stationary environments where the classification task changes during the operation of the ensemble and the concept of ”forgetting” is discussed.
Change with Delayed Labeling: When is it Detectable?
Indre liobaite
- 13 Dec 2010
TL;DR: This analytical study looks at the space of changes from probabilistic perspective to analyze, what changes are detectable without seeing the labels and what are not, and proposes a computationally friendly detection technique, which monitors a stream of classifier outputs.