Learning in the Presence of Concept Drift and Hidden Contexts

By Gerhard Widmer, Miroslav Kubat. 1996

How researchers see this paper

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

Citations in the corpus

From amateurs to connoisseurs: modeling the evolution of user expertise through online reviews (2013, background): “Some early works that deal with temporal dynamics do so in terms of concept drift [23, 37, 39].”
Domain Generalization via Invariant Feature Representation (2013, background): “In this case, the data exhibits covariate (or dataset) shift (Widmer and Kurat, 1996; Quionero-Candela et al., 2009; Bickel et al., 2009): although the marginal distributions PX on cell attributes vary due to biological or technical variations, the functional relationship P(Y |X) across different do”
Mind the Gap: Assessing Temporal Generalization in Neural Language Models (2021, background): “Detecting changes in data streams, also known as concept drift, has a long history in machine learning (Widmer and Kubat, 1996; Kifer et al., 2004; Baena-Garcıa et al., 2006; Dries and Rückert, 2009; Lu et al., 2019).”
WILDS: A Benchmark of in-the-Wild Distribution Shifts (2020, neutral): “Second, many distribution shifts in medicine arise from concept drifts, in which the relationship between the input and the label changes, for example due to changes in clinical procedures and the definition of the label (Widmer and Kubat, 1996; Beyene et al., 2015; Futoma et al., 2020).”
Temporal recommendation on graphs via long- and short-term preference fusion (2010, background): “Unlike traditional concept drifting mining [20, 24] that focuses primarily on global pattern shifting over time, temporal recommendation focuses on local recommendation models for each user.”
Classifier Technology and the Illusion of Progress (2006, unknown): “2), Widmer and Kubat [46] and Lane and Brodley [30].”
Learned lessons in credit card fraud detection from a practitioner perspective (2014, methodology): “It is well known however that an open issue is how to manage unbalanced class sizes since the legitimate transactions generally far outnumber the fraudulent ones.”
Deep Learning for Anomaly Detection in Time-Series Data: Review, Analysis, and Guidelines (2021, background): “This phenomenon goes by many names in literature, the most common of which is concept drift [21].”
Scalable and efficient multi-label classification for evolving data streams (2012, methodology): “A simple approach to the data stream problem is to train a batch-learning classifier on batches of new examples that replace classifiers built from earlier batches over time (Widmer and Kubat 1996; Qu et al. 2009).”
Algorithm fairness in artificial intelligence for medicine and healthcare (2023, background): “This is known as concept shift or concept drift 181–186 , and it involves a change in the conditional distributions P ( X | Y ) or P ( Y | X ) while the marginal distributions P ( X ) and P ( Y ) remain unchanged.”
Online Machine Learning in Big Data Streams (2018, methodology): “ng of the window. Next, a corresponding number of examples are removed from the end of the window, and the learner is reapplied, as in early research on concept drift in learning from continuous data [228, 119]. Finally, as a combination with online learning methods, the batch trained model can be i”
A Review of Domain Adaptation without Target Labels (2019, background): “Concept shift, also known as concept drift or conditional shift, will require observations of labeled data in both domains and is therefore out of our scope [86].”
Transferring Instances for Model-Based Reinforcement Learning (2008, background): “Note that TL is related to, but distinct from, the concept drift [8] and multi-task learning [9, 10] paradigms.”
Characterizing concept drift (2015, background): “To this end there has been much work on identifying types of concept drift [Widmer and Kubat, 1996, Gama and Rodrigues, 2009, Zliobaite, 2010, Hoens et al., 2012, Moreno-Torres et al., 2012, Gama et al., 2014].”
Applying lazy learning algorithms to tackle concept drift in spam filtering (2007, methodology): “Some algorithms use a time window of ﬁxed size (Kubat, 1989), while others use heuristics to adjust the window size to the current extent of concept drift (Klinkenberg, 2004; Widmer & Kubat, 1996).”
Concept drift detection via competence models (2014, methodology): “The issue of concept drift refers to the change of distribution underlying the data [4, 8].”
Collaborative filtering with temporal dynamics (2010, methodology): “[15, 25 ]) used techniques like adjusted and decayed weights of past instances or using a sliding time window.”