Knowledge-maximized ensemble algorithm for different types of concept drift

Ren, Siqi; Liao, Bo; Zhu, Wen; Li, Keqin

doi:10.1016/j.ins.2017.11.046

Cited by 62 publications

(35 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…To compare various algorithms and to show if there exist significant differences among them, it is essential to give statistical test support. This paper investigates the usage of Friedman and Wilcoxon tests for machine learning methods [4,6,14]. The null hypothesis for the experimental design suggests that there exists no significant difference between the prediction performances of the algorithms tested.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Dynamically updated diversified ensemble-based approach for handling concept drift

Goel¹,

Batra²

2020

Turk J Elec Eng & Comp Sci

View full text Add to dashboard Cite

Concept drift is the phenomenon where underlying data distribution changes over time unexpectedly.Examining such drifts and getting insight into the executing processes at that instance of time is a big challenge.Prediction models should be capable of handling drifts in scenarios where statistical properties show abrupt changes.Various strategies exist in the literature to deal with such challenging scenarios but the majority of them are limited to the identification of a particular kind of drift pattern. The proposed approach uses online drift detection in a diversified adaptive setting with pruning techniques to formulate a concept drift handling approach, named ensemblebased online diversified drift detection (En-ODDD), with an aim to identify the majority of drifts including abrupt, gradual, recurring, mixed, etc. in a single model. En-ODDD is equipped with a dynamically updated ensemble to speed up the adaptability to changing distributions. Unlike prevalent approaches, which do not consider correlations between experts, En-ODDD entails experts using varying randomized subsets of input data. Different levels of sampling having been applied for diversity generation to promote generalization. Prediction accuracy has been used to evaluate the effectiveness of the proposed approach using Massive Online Analysis software and compared with ten state-of-theart algorithms. Experimental results on fifteen benchmark datasets (artificial and real-world) having up to one million instances depict that En-ODDD outperforms the existing approaches irrespective of nature of drift.

show abstract

Section: Discussionmentioning

confidence: 99%

“…In block-based techniques, data are processed in the form of fixed-size or variably sized batches [3,4], whereas online approaches analyze instances on the go [5,6]. Both of them can use either an explicit drift detection mechanism [7][8][9] or implicit adaptive strategy to handle evolving data distribution.…”

Section: Introductionmentioning

confidence: 99%

Dynamically updated diversified ensemble-based approach for handling concept drift

Goel¹,

Batra²

2020

Turk J Elec Eng & Comp Sci

View full text Add to dashboard Cite

show abstract

“…Knowledge Maximized Ensemble (KME) uses a combination of off-the-shelf and created drift detectors to recognize various forms of drift simultaneously. Models are updated when enough training data is collected and removed if they perform poorly [18].…”

Section: Concept Drift Adaptationmentioning

confidence: 99%

Concept Drift Adaptive Physical Event Detection for Social Media Streams

Suprem

Musaev

2019

Services – SERVICES 2019

View full text Add to dashboard Cite

Event detection has long been the domain of physical sensors operating in a static dataset assumption. The prevalence of social media and web access has led to the emergence of social, or human sensors who report on events globally. This warrants development of event detectors that can take advantage of the truly dense and high spatial and temporal resolution data provided by more than 3 billion social users. The phenomenon of concept drift, which causes terms and signals associated with a topic to change over time, renders static machine learning ineffective. Towards this end, we present an application for physical event detection on social sensors that improves traditional physical event detection with concept drift adaptation. Our approach continuously updates its machine learning classifiers automatically, without the need for human intervention. It integrates data from heterogeneous sources and is designed to handle weak-signal events (landslides, wildfires) with around ten posts per event in addition to large-signal events (hurricanes, earthquakes) with hundreds of thousands of posts per event. We demonstrate a landslide detector on our application that detects almost 350% more landslides compared to static approaches. Our application has high performance: using classifiers trained in 2014, achieving event detection accuracy of 0.988, compared to 0.762 for static approaches.

show abstract

“…Such assumption typically means that data used to train the predictive models can reflect the probability distribution of the problem. However, this assumption is often violated in real-world applications (Gállego et al 2017;Ren et al 2018). For many reasons, the data distribution in real-world applications is often not stable and tends to change with time (Tsymbal 2004;Zliobaite et al 2016).…”

Section: Introductionmentioning

confidence: 99%

An ensemble based on neural networks with random weights for online data stream regression

et al. 2019

View full text Add to dashboard Cite

Most information sources in the current technological world are generating data sequentially and rapidly, in the form of data streams. The evolving nature of processes may often cause changes in data distribution, also known as concept drift, which is difficult to detect and causes loss of accuracy in supervised learning algorithms. As a consequence, online machine learning algorithms that are able to update actively according to possible changes in the data distribution are required. Although many strategies have been developed to tackle this problem, most of them are designed for classification problems. Therefore, in the domain of regression problems, there is a need for the development of accurate algorithms with dynamic updating mechanisms that can operate in a computational time compatible with today's demanding market. In this article, the authors propose a new bagging ensemble approach based on neural network with random weights for online data stream regression. The proposed method improves the data prediction accuracy as well as minimises the required computational time compared to a recent algorithm for online data stream regression from literature. The experiments are carried out using four synthetic datasets to evaluate the algorithm's response to concept drift, along with four benchmark datasets from different industries. The results indicate improvement in data prediction accuracy, effectiveness in handling concept drift, and much faster updating times compared to the existing available approach. Additionally, the use of design of experiments as an effective tool for hyperparameter tuning is demonstrated. Keywords Ensembles Á Data stream regression Á Neural networks with random weights Á Hyperparameter adjustment Communicated by V. Loia.

show abstract

Knowledge-maximized ensemble algorithm for different types of concept drift

Cited by 62 publications

References 18 publications

Dynamically updated diversified ensemble-based approach for handling concept drift

Dynamically updated diversified ensemble-based approach for handling concept drift

Concept Drift Adaptive Physical Event Detection for Social Media Streams

An ensemble based on neural networks with random weights for online data stream regression

Contact Info

Product

Resources

About