Parameter estimation in abruptly changing dynamic environments using stochastic learning weak estimator

Hammer, Hugo Lewi; Yazidi, Anis

doi:10.1007/s10489-018-1205-3

Cited by 7 publications

(3 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, a disadvantage of window-based approaches is that every sample in the window is weighted equally. Another computationally efficient approach that addresses this issue relies on tracking the current prediction error using the exponentially weighted moving average [12,40].…”

Section: Concept Driftmentioning

confidence: 99%

See 1 more Smart Citation

Efficient quantile tracking using an oracle

et al. 2022

Self Cite

View full text Add to dashboard Cite

Concept drift is a well-known issue that arises when working with data streams. In this paper, we present a procedure that allows a quantile tracking procedure to cope with concept drift. We suggest using expected quantile loss, a popular loss function in quantile regression, to monitor the quantile tracking error, which, in turn, is used to efficiently adapt to concept drift. The suggested procedures adapt efficiently to concept drift, and the tracking performance is close to theoretically optimal. The procedures were further applied to three real-life streaming data sets related to Twitter event detection, activity recognition, and stock trading. The results show that the procedures are efficient at adapting to concept drift, thereby documenting the real-world applicability of the procedures. We further used asymptotic theory from statistics to show the appealing theoretical property that, if the data stream distribution is stationary over time, the procedures converge to the true quantile.

show abstract

Section: Concept Driftmentioning

confidence: 99%

“…To estimate the current expected quantile loss, we use the exponentially weighted moving average [12] EQL…”

Section: Adaptive Quantile Tracking Under Concept Driftmentioning

confidence: 99%

Efficient quantile tracking using an oracle

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…Theoretical extensions of the SPL. In Hammer and Yazidi (2018), Hammer and Yazidi presented a method by which the SLWE could detect abrupt changes in data streams. To achieve this goal, two parallel weak estimators were run in parallel.…”

Section: Unique Applications Of the Slwe In Non-stationary Environmentsmentioning

confidence: 99%

On utilizing weak estimators to achieve the online classification of data streams

Tavasoli

Oommen

Yazidi

2019

Engineering Applications of Artificial Intelligence

Self Cite

View full text Add to dashboard Cite

Classification, typically, deals with unique and distinct training and testing phases. This paper pioneers the concept when these phases are not so clearly well-defined. More specifically, we consider the case where the testing patterns can subsequently be considered as training patterns. The paradigm is further complicated because we assume that the class-conditional distributions of the features/classes are non-stationary, as in the case of most real-world applications. Specifically, we consider the model where the training phase is nonstationary and that it is, further, interleaved with the testing, and where it can be done online and in a real-time manner.We propose a novel online classifier for complex data streams which are generated from non-stationary stochastic properties. Instead of using a single training model with ''counters'' that maintain important data statistics, our online classifier scheme provides a real-time self-adjusting learning model. The learning model utilizes the multiplication-based update algorithm of the Stochastic Learning Weak Estimator (SLWE) at each time instant as a new labeled instance arrives. In this way, the data statistics are updated every time a new element is seen, without requiring that we have to rebuild the model when changes occur in the data distributions. Finally, and most importantly, the model operates with the understanding that the correct classes of previously-classified patterns become available at a later juncture subsequent to some time instances. This forces us to update the training set, the training model and the class conditional distributions as the testing proceeds.The results from rigorous empirical analysis on two-dimensional/multi-dimensional and binomial/ multinomial distributions are remarkable. We also report some results on two real-life datasets adapted to this model of computation, demonstrating the advantages of the novel scheme for both binomial and multinomial non-stationary distributions.

show abstract

On solving the SPL problem using the concept of probability flux

2019

Self Cite

View full text Add to dashboard Cite

The Stochastic Point Location (SPL) problem [20] is a fundamental learning problem that has recently found a lot of research attention. SPL can be summarized as searching for an unknown point in an interval under faulty feedback. The search is performed via a Learning Mechanism (LM) (algorithm) that interacts with a stochastic Environment which in turn informs it about the direction of the search. Since the Environment is stochastic, the guidance for directions could be faulty. The first solution to the SPL problem, which was pioneered two decades ago by Oommen, relies on discretizing the search interval and performing a controlled random walk on it. The state of the random walk at each step is considered to be the estimation of the point location. The convergence of the latter simplistic estimation strategy is proved for an infinite resolution, i.e., infinite memory. However, this strategy yields rather poor accuracy for low discretization resolutions. In this paper, we present two major contributions to the SPL problem. First, we demonstrate that the estimation of the point location can significantly be improved by resorting to the concept of mutual probability flux between neighboring states along the line. Second, we are able to accurately track the position of the optimal point and simultaneously show a method by which we can estimate the error probability characterizing the Environment. Interestingly, learning this error probability of the Environment takes place in tandem with the unknown location estimation. We present and analyze several experiments discussing the weaknesses and strengths of the different methods.

show abstract

Parameter estimation in abruptly changing dynamic environments using stochastic learning weak estimator

Cited by 7 publications

References 41 publications

Efficient quantile tracking using an oracle

Efficient quantile tracking using an oracle

On utilizing weak estimators to achieve the online classification of data streams

On solving the SPL problem using the concept of probability flux

Contact Info

Product

Resources

About