A survey on feature drift adaptation: Definition, benchmark, challenges and future directions

Barddal, Jean Paul; Gomes, Heitor Murilo; Enembreck, Fabrício; Pfahringer, Bernhard

doi:10.1016/j.jss.2016.07.005

Cited by 89 publications

(50 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Heterogeneous Ensemble with Feature drifT for Data Streams (HEFT-Stream) [23] is an online classifier that incorporates feature selection by applying the Fast Correlation Based Filter (FCBF) [26] algorithm that dynamically updates the relevant feature subsets for data streams. This is beneficial because non-stationary environments may present feature drift [23] [41]. In high-dimensional datasets, not all features are significant for training a classifier and the relevance of a feature may grow or shrink over time.…”

Section: Heterogeneous Ensemblesmentioning

confidence: 99%

A heterogeneous online learning ensemble for non-stationary environments

Idrees

Minku

Stahl

et al. 2020

Knowledge-Based Systems

View full text Add to dashboard Cite

Learning in non-stationary environments is a challenging task which requires the updating of predictive models to deal with changes in the underlying probability distribution of the problem, i.e., dealing with concept drift. Most work in this area is concerned with updating the learning system so that it can quickly recover from concept drift, while little work has been dedicated to investigating what type of predictive model is most suitable at any given time. This paper aims to investigate the benefits of online model selection for predictive modelling in nonstationary environments. A novel heterogeneous ensemble approach is proposed to intelligently switch between different types of base models in an ensemble to increase the predictive performance of online learning in nonstationary environments. This approach is Heterogeneous Dynamic Weighted Majority (HDWM). It makes use of "seed" learners of different types to maintain ensemble diversity, overcoming problems of existing dynamic ensembles that may undergo loss of diversity due to the exclusion of base learners. The algorithm has been evaluated on artificial and real-world data streams against existing well-known approaches such as a heterogeneous Weighted Majority Algorithm (WMA) and a homogeneous Dynamic Weighted Majority (DWM). The results show that HDWM performed significantly better than WMA in non-stationary environments. Also, when recurring concept drifts were present, the predictive performance of HDWM showed an improvement over DWM.

show abstract

Section: Heterogeneous Ensemblesmentioning

confidence: 99%

A heterogeneous online learning ensemble for non-stationary environments

Idrees

Minku

Stahl

et al. 2020

Knowledge-Based Systems

View full text Add to dashboard Cite

show abstract

“…Recently, the works of [16] and [17] surveyed and evaluated different approaches to tackle this problem. According to these studies, incremental decision trees [18] and its variants [19] are the best performing approaches.…”

Section: Related Workmentioning

confidence: 99%

“…Regarding synthetic experiments, AGR represents the AGRAWAL [43] generator and AN is the Asset Negotiation generator [44]. BG1, BG2 and BG3 are synthetic generators based on binary features proposed in [45] that were recently used to synthetize feature drifts in [17]. Finally, the Random Tree Generator (RTG) was used to create more complex concepts (where the number of relevant features is bigger), while SEA [46] concepts depend on only 2 features.…”

Section: Experimental Protocolmentioning

confidence: 99%

“…If we analyze the classification and selection accuracy rates, depicted in Figures 5a and 5b, respectively; we observe that these experiments result in rates that are much lower than the rest. The explanation is that these concepts are much more complex than the others, as BG3 is a XOR-like classification problem [45], and RTG has complex interactions between the features [17]. In practice, the Selection Accuracy rates obtained in the BG3 and RTG experiments are below the expected baseline of 0.7.…”

Section: Synthetic Experimentsmentioning

confidence: 99%

See 1 more Smart Citation

Boosting decision stumps for dynamic feature selection on data streams

Barddal

Enembreck

Gomes

et al. 2019

Information Systems

Self Cite

View full text Add to dashboard Cite

Feature selection targets the identification of which features of a dataset are relevant to the learning task. It is also widely known and used to improve computation times, reduce computation requirements, and to decrease the impact of the curse of dimensionality and enhancing the generalization rates of classifiers. In data streams, classifiers shall benefit from all the items above, but more importantly, from the fact that the relevant subset of features may drift over time. In this paper, we propose a novel dynamic feature selection method for data streams called Adaptive Boosting for Feature Selection (ABFS). ABFS chains decision stumps and drift detectors, and as a result, identifies which features are relevant to the learning task as the stream progresses with reasonable success. In addition to our proposed algorithm, we bring feature selection-specific metrics from batch learning to streaming scenarios. Next, we evaluate ABFS according to these metrics in both synthetic and real-world scenarios. As a result, ABFS improves the classification rates of different types of learners and eventually enhances computational re

show abstract

“…There are different types of concept drift detection mechanisms for handling gradual or abrupt changes, blips or recurring drifts [ 24 , 26 , 27 , 28 , 52 ] that can be used to deal with changes in the market behavioral structure [ 53 ]. As opposed to stationary data distributions, where the error rate of the learning algorithm will decrease when the number of examples increases, the presence of changes affects the learning model continuously [ 54 ].…”

Section: Related Workmentioning

confidence: 99%

Incremental Market Behavior Classification in Presence of Recurring Concepts

Suárez‐Cetrulo

Cervantes

Quintana

2019

Entropy

View full text Add to dashboard Cite

In recent years, the problem of concept drift has gained importance in the financial domain. The succession of manias, panics and crashes have stressed the non-stationary nature and the likelihood of drastic structural or concept changes in the markets. Traditional systems are unable or slow to adapt to these changes. Ensemble-based systems are widely known for their good results predicting both cyclic and non-stationary data such as stock prices. In this work, we propose RCARF (Recurring Concepts Adaptive Random Forests), an ensemble tree-based online classifier that handles recurring concepts explicitly. The algorithm extends the capabilities of a version of Random Forest for evolving data streams, adding on top a mechanism to store and handle a shared collection of inactive trees, called concept history, which holds memories of the way market operators reacted in similar circumstances. This works in conjunction with a decision strategy that reacts to drift by replacing active trees with the best available alternative: either a previously stored tree from the concept history or a newly trained background tree. Both mechanisms are designed to provide fast reaction times and are thus applicable to high-frequency data. The experimental validation of the algorithm is based on the prediction of price movement directions one second ahead in the SPDR (Standard & Poor’s Depositary Receipts) S&P 500 Exchange-Traded Fund. RCARF is benchmarked against other popular methods from the incremental online machine learning literature and is able to achieve competitive results.

show abstract

A survey on feature drift adaptation: Definition, benchmark, challenges and future directions

Cited by 89 publications

References 48 publications

A heterogeneous online learning ensemble for non-stationary environments

A heterogeneous online learning ensemble for non-stationary environments

Boosting decision stumps for dynamic feature selection on data streams

Incremental Market Behavior Classification in Presence of Recurring Concepts

Contact Info

Product

Resources

About