Learning from Time-Changing Data with Adaptive Windowing

By Albert Bifet, Ricard Gavaldà. SDM, 2007

1,868 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Signal Processing

View Paper: https://www.semanticscholar.org/paper/d7f8d7b89593c5333eb174b2411bf004f5a91f7d

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 Signal Processing

Citations in the corpus

Automatic Model Monitoring for Data Streams (2019, methodology): “two most frequently used methods to detect concept drift are ADWIN [2] and the Page-Hinkley test [10].”
Adaptive Deep Models for Incremental Learning: Considering Capacity Scalability and Sustainability (2019, methodology): “Considering IADM is related to the deep incremental learning with concept drift, several state-of-the-art methods are compared, i.e., DNN-SGD, Adwin [3], ODL [31].”
MOA: Massive Online Analysis, a Framework for Stream Classification and Clustering (2010, neutral): “ADWIN keeps a variable-length window of recently seen items, with the property that the window has the maximal length statistically consistent with the hypothesis " there has been no change in the average value inside the window " .”
Learning under Concept Drift: an Overview (2010, background): “Change detection can be based on monitoring the raw data [13, 119], the parameters of the learners [143] or the outputs (error) of the learners [55, 5, 114].”
Minimax Classification under Concept Drift with Multidimensional Adaptation and Performance Guarantees (2022, neutral): “…learning techniques account for a scalar rate of change by means of a carefully chosen parameter such as a learning rate (Kivinen et al., 2004; Orabona et al., 2008), forgetting factor (Klinkenberg, 2004; Pavlidis et al., 2011), or window size (Bifet & Gavalda, 2007; Nguyen et al., 2018).”
Reacting to Different Types of Concept Drift: The Accuracy Updated Ensemble Algorithm (2014, background): “For this reason, more dynamic ways of modeling the forgetting process, such as heuristic adjusting of the window size [19] or decay functions [20], have been proposed.”
Credit Card Fraud Detection: A Realistic Modeling and a Novel Learning Strategy (2018, background): “Active approaches [4], [9], [34], [50], [60] use a change-”
Ensemble learning for data stream analysis: A survey (2017, background): “In this algorithm [16] a window of incoming examples grows until identifying a change in the average value inside the window.”
A survey on concept drift adaptation (2014, methodology): “The ADaptive sliding WINdow or ADWIN [Bifet and Gavalda 2006; 2007] is another change detector using a detection window.”
Extremely Fast Decision Tree (2018, methodology): “HAT uses an error estimator, such as ADWIN [5] at each node to determine whether the prediction error due to a recent sequence of examples is significantly greater than the prediction error from a longer historical sequence so it can respond to drift.”
Scalable and efficient multi-label classification for evolving data streams (2012, background): “A formal and quantitative statement of these two points (in the form of a theorem) appears in Bifet and Gavaldà (2007).”
Online and Non-Parametric Drift Detection Methods Based on Hoeffding’s Bounds (2015, methodology): “In the following subsections we mention relevant algorithms used in online learning, which fall in two main categories: window strategies and weightedinstance approaches.”
Concept drift detection via competence models (2014, methodology): “In this sense, we intentionally omit methods that are designed for one-dimensional data [37, 38].”
Leveraging Bagging for Evolving Data Streams (2010, background): “ADWIN is parameter- and assumption-free in the sense that it automatically detects and adapts to the current rate of change.”
Detecting concept drift in data streams using model explanation (2018, methodology): “R I P Representative algorithms for drift detection using windowing strategies are FLORA (Widmer & Kubat, 1996), ADWIN (Bifet & Gavalda, 2007), CUSUM (Page, 1954), Page-Hinkley test (PH) (Page, 1954), Statistical Process Control (SPC) (Gama, 2010), RCD (Gon¸calves Jr & Barros, 2012) and others (Hoen”
On evaluating stream learning algorithms (2012, methodology): “…growing research ﬁeld and several methods capable of dealing with change detection have been presented in the literature (Hulten et al., 2001; Basseville and Nikiforov, 1993; Kifer et al., 2004; Widmer and Kubat, 1996; Klinkenberg, 2004; Koychev, 2000; Gama et al., 2004; Bifet and Gavald`a,…”