Incremental Learning of Concept Drift in Nonstationary Environments

By Ryan Elwell, Robi Polikar. IEEE Transactions on Neural Networks, 2011

930 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Management Science and Operations Research

View Paper: https://www.semanticscholar.org/paper/ed698fd7a4ca267a0279cabdf412db7f6632009f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Management Science and Operations Research

Citations in the corpus

Road Traffic Forecasting: Recent Advances and New Challenges (2018, background): “Incremental learning implies any instance of data can only be used once for training [126] so older instances become less relevant when the algorithm evolves.”
Incremental on-line learning: A review and comparison of state of the art algorithms (2018, background): “Various algorithms have been published explicitly tackling this challenge [74, 75, 76].”
Concept drift detection via competence models (2014, background): “Elwell and Polikar [24] presented an ensemble learning algorithm for non-stationary environments (Learn++.”
A Survey on Ensemble Learning for Data Stream Classification (2017, background): “…drifting concepts [Kolter et al. 2003; Bifet et al. 2009; Bifet et al. 2010a; Gomes and Enembreck 2014; Dongre and Malik 2013; Deck-ert 2011; Elwell and Polikar 2011], noisy data [Zhu et al. 2004; Zhang et al. 2011b], recurrent concepts [Nishida et al. 2005; Ramamurthy and Bhatnagar 2007;…”
Credit Card Fraud Detection: A Realistic Modeling and a Novel Learning Strategy (2018, methodology): “Ensemble methods [23], [30], [43], [61], [72] and classifiers trained over a sliding window of the recent supervised samples (like STAGGER [57] and FLORA [67]) are probably the most extensively investigated passive solutions.”
Online Learning: A Comprehensive Survey (2018, methodology): “They just adapt some batch learning algorithms to work efficiently for seqential learning environments, such as the series of Learning++ algorithms and their variants [213, 215].”
Incremental Evolving Domain Adaptation (2016, methodology): “We use the same eight features as [38].”
Cost-Effective Incremental Deep Model: Matching Model Capacity With the Least Sampling (2023, neutral): “…on one synthetic ( Hyperplane [52]) and two constructed image incremental datasets ( Incremental MNIST [53], Incremental CIFAR10 [54]), then give the analysis on three real-world datasets, i.e., action recognition ( Incremental UCF101 ) [55], Weather (Wea) [56] and Electricity (Elec) [57].”
From Concept Drift to Model Degradation: An Overview on Performance-Aware Drift Detectors (2022, background): “Similarly, concept deletion refers to the disappearance of classes in the problem [20], Figure 7(b).”
Error-Driven Incremental Learning in Deep Convolutional Neural Network for Large-Scale Image Classification (2014, background): “The ﬁrst is concept drift in the training dataﬂow, and therefore the classiﬁer learns in a non-stationary environment [5, 24].”
Learning in Nonstationary Environments: A Survey (2015, methodology): “tionary environments) [10], whose pseudocode is shown in Figure 5.”
Distribution-Free One-Pass Learning (2017, methodology): “NSE [39], and (f) OAUE [50], where (b)-(f) are all ensemble style approaches which dynamically adapt themodels based”
Incremental learning algorithms and applications (2016, methodology): “Ensemble methods have proved particularly useful when dealing with concept drift, with a few popular models ranging from incremental random forests [105], ensembles of bipartite graph classifiers [13], up to advanced weighting schemes suitable for different types of concept drift and recurring conce”
Reacting to Different Types of Concept Drift: The Accuracy Updated Ensemble Algorithm (2014, methodology): “results in data stream mining papers [26], [31].”
Learning with Feature and Distribution Evolvable Streams (2020, background): “…falls into the ensemble based approaches (Zhou, 2012; Gomes et al., 2017), where they adaptively add or drop base classiﬁers learning from historical data and dynamically adjust weights when dealing with new coming data items (Kolter & Maloof, 2005; Elwell & Polikar, 2011; Zhao et al., 2020).”
IncDet: In Defense of Elastic Weight Consolidation for Incremental Object Detection (2020, background): “NSE [31] and its class imbalance variant [32] study incremental learning of concept drift in nonstationary environments.”
A survey on data preprocessing for data stream mining: Current status and future directions (2017, background): “Two main approaches here assume a changing lineup of the ensemble [37, 38, 39] or updating base classifiers [40, 41].”