Early Classification of Time Series as a Non Myopic Sequential Decision Making Problem

Dachraoui, Asma; Bondu, Alexis; Cornuéjols, Antoine

doi:10.1007/978-3-319-23528-8_27

Cited by 42 publications

(81 citation statements)

References 9 publications

(16 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For example, in [4], the authors formalized the problem of early classification of time series as a sequential decision problem involving two costs: quality and delay of the prediction. Their method also provides the estimated time for classification, that is, how much of the remaining time series is needed to classify.…”

Section: Related Workmentioning

confidence: 99%

Learning When to Classify for Early Text Classification

Loyola

Errecalde

Gómez

2018

Communications in Computer and Information Science

View full text Add to dashboard Cite

Abstract. The problem of classification in supervised learning is a widely studied one. Nonetheless, there are scenarios that received little attention despite its applicability. One of such scenarios is early text classification, where one needs to know the category of a document as soon as possible. The importance of this variant of the classification problem is evident in tasks like sexual predator detection, where one wants to identify an offender as early as possible. This paper presents a framework for early text classification which highlights the two main pieces involved in this problem: classification with partial information and deciding the moment of classification. In this context, a novel approach that learns the second component (when classify) and an adaptation of a temporal measurement for multi-class problems are introduced. Results with a classical text classification corpus in comparison against a model that reads the entire documents confirm the feasibility of our approach.

show abstract

Section: Related Workmentioning

confidence: 99%

Learning When to Classify for Early Text Classification

Loyola

Errecalde

Gómez

2018

Communications in Computer and Information Science

View full text Add to dashboard Cite

show abstract

“…Recently, Dachraoui et al [3] introduced a framework for cost-aware early classification of time series that is dedicated to optimizing a trade-o↵ between the accuracy of the prediction and the time at which it is performed. This framework hence involves both costs defined above: a misclassification cost and a cost of delaying the prediction.…”

Section: Introductionmentioning

confidence: 99%

“…This framework hence involves both costs defined above: a misclassification cost and a cost of delaying the prediction. The authors of [3] derive an algorithm that decides, at test time, whether the decision should be made at a given time instant t or if more data should be waited for (based on both costs defined above). In addition, the method of [3] has two other interesting properties : it is adaptive and nonmyopic.…”

Section: Introductionmentioning

confidence: 99%

“…The authors of [3] derive an algorithm that decides, at test time, whether the decision should be made at a given time instant t or if more data should be waited for (based on both costs defined above). In addition, the method of [3] has two other interesting properties : it is adaptive and nonmyopic. It is adaptive in the sense that the optimal time of decision depends on the time series to be classified.…”

Section: Introductionmentioning

confidence: 99%

“…In this paper, we propose two new early-classification schemes built upon the framework defined in [3]. This framework is based on approximating an expected cost through clustering, which tends to bring vagueness in the process.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Cost-Aware Early Classification of Time Series

Tavenard

Malinowski

2016

Machine Learning and Knowledge Discovery in Databases

View full text Add to dashboard Cite

Abstract. In time series classification, two antagonist notions are at stake. On the one hand, in most cases, the sooner the time series is classified, the more rewarding. On the other hand, an early classification is more likely to be erroneous. Most of the early classification methods have been designed to take a decision as soon as su cient level of reliability is reached. However, in many applications, delaying the decision with no guarantee that the reliability threshold will be met in the future can be costly. Recently, a framework dedicated to optimizing a trade-o↵ between classification accuracy and the cost of delaying the decision was proposed, together with an algorithm that decides online the optimal time instant to classify an incoming time series. On top of this framework, we build in this paper two di↵erent early classification algorithms that optimize a trade-o↵ between decision accuracy and the cost of delaying the decision. These algorithms are non-myopic in the sense that, even when classification is delayed, they can provide an estimate of when the optimal classification time is likely to occur. Our experiments on real datasets demonstrate that the proposed approaches are more robust than existing methods.

show abstract

A novel approach for early malware detection

Sharma

Singh

2020

Trans Emerging Tel Tech

View full text Add to dashboard Cite

Early classification of time series is valuable in many real‐world applications such as early disease prediction, early disaster prediction, and patient monitoring where data are generated over time. The main objective of early classification is to provide a reliable class prediction earliest in time. In general, whenever the early prediction time improves, the prediction accuracy decreases. Thus, the trade‐off between earliness and accuracy needs to be addressed. In this article, we proposed an optimization‐based early classification model for time series data using early stopping rules (ESRs) and a series of probabilistic classifiers. ESRs are developed through particle swarm optimization by minimizing the well‐defined cost function that considers the missclassification cost and delaying decision cost simultaneously. The experimental results on 30 standard datasets demonstrate good performance for early classification in comparison to state of the art methods. Also, the proposed model is tested for early malware detection on a real dataset and shows decent performance by balancing the accuracy and earliness.

show abstract

Early Classification of Time Series as a Non Myopic Sequential Decision Making Problem

Cited by 42 publications

References 9 publications

Learning When to Classify for Early Text Classification

Learning When to Classify for Early Text Classification

Cost-Aware Early Classification of Time Series

A novel approach for early malware detection

Contact Info

Product

Resources

About