The Impact of Diversity on Online Ensemble Learning in the Presence of Concept Drift

By Leandro L. Minku, Allan P. White, Xin Yao. IEEE Transactions on Knowledge and Data Engineering, 2010

477 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Management Science and Operations Research

View Paper: https://www.semanticscholar.org/paper/53b622d632096ab8060aaf87c9d2f6aa68951e40

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

Online Machine Learning in Big Data Streams (2018, background): “ast, do not aim to detect changes but rather maintain the most accurate models at each time step. Evolving learners are method-speciﬁc, most of them based on ensemble methods [225, 139]. A few papers [168, 171] give overviews of different types of environmental changes and concept drifts based on sp”
Incremental Learning of Concept Drift in Nonstationary Environments (2011, background): “…active scaffolding enhances learning by complexity reduction [ ↔ preprocessing, feature selection], prob-lematizing [7], i.e., ﬁnding conﬂicts between current and prior knowledge [ ↔ drift detecting], and fading [7], [8], preventing redundancy when an environment has been learned [ ↔ pruning].”
Ensemble learning for data stream analysis: A survey (2017, background): “DDD ensemble [128] or other generalizations of online bagging such as [18].”
From Concept Drift to Model Degradation: An Overview on Performance-Aware Drift Detectors (2022, background): “Intersected concept drift occurs when only a subspace of the data samples changes their target classes after the drift occurrence [53], which is also referred to as subconcept drift [48], Figure 6(e).”
An Overview of Concept Drift Applications (2016, unknown): “[55]).”
Reacting to Different Types of Concept Drift: The Accuracy Updated Ensemble Algorithm (2014, background): “Another type of gradual drift, which we will refer to as incremental, includes more than two sources, however the difference between them is small and the change is noticed in a longer period of time [5], [16].”
Ensemble learning: A survey (2018, background): “Another work examined how different diversity levels of ensembles affect their generalization for new concepts (Minku, White, & Yao, 2010).”
Learning under Concept Drift: an Overview (2010, background): “rgue, that assumption about the change types is absolutely needed for designing adaptivity strategies. Recently there has been an attempt to categorize change types into mutually exclusive categories [108] based on number of reoccurences, severity, speed and predictability. In principle the proposed”
Distribution-Free One-Pass Learning (2017, background): “Note that although there are plenty of studies about learning with distribution change, most approaches are developed based on the use of sliding windows [18], [19], [20] or ensemble strategies [4], [5], [21], [22].”
A Survey on Ensemble Learning for Data Stream Classification (2017, methodology): “[2010] to conduct diversity-related experiments in evolving data streams for mainly two reasons: (1) “it does not present any specific behaviour to handle concept drift” [Minku et al. 2010], and (2) diversity in online bagging is “controlled” through a single parameter λ.”
Practical Ensemble Classification Error Bounds for Different Operating Points (2013, background): “For example, consider an analyst at an information technology (IT) outsourcing provider who must label service tickets into categories such as network, operating system, or hardware to analyze what problems are frequently occurring at client installations [5].”
Ensemble deep learning: A review (2021, methodology): “Other examples that employed online ensemble learning models were used to deal with the presence of concept drift [177], power load forecasting [178, 179], myoelectric prosthetic hands surface electromyogram characteristics [180], etc.”
Learning in Nonstationary Environments: A Survey (2015, background): “Recent work has shown that using both high and low diversity ensembles can be beneficial for tracking different rates of drift in the data stream [85], [86].”
DDD: A New Ensemble Approach for Dealing with Concept Drift (2012, methodology): “For a detailed explanation of different types of drift, we recommend [9].”