Early Fault Classification in Dynamic Systems Using Case-Based Reasoning

Bregón, Aníbal; Simón, M. Aránzazu; Rodrı́guez, Juan J.; Alonso, Carlos J.; Pulido, Belarmino; Moro, Isaac

doi:10.1007/11881216_23

Cited by 27 publications

(32 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, Evans et al [10] show that the monitoring of patients and early identification of physiologic deterioration can be used to raise alerts and prevent crises in hospitalized patients. Also, Ghalwash et al [11] mention early stock crisis identification; Bregón et al [12] apply early classification to classify different types of faults in a simulated industrial plant; Hatami and Chira [13] attempt to classify a set of different odors as early as possible by using odor signals obtained from a set of sensors with the aim of identifying chemical leaks. Finally, in Mori et al [14], an early classification approach is applied to detect and identify bird songs as early as possible, with the objective of saving memory and battery life of the recording devices.…”

Section: Introductionmentioning

confidence: 99%

Early Classification of Time Series by Simultaneously Optimizing the Accuracy and Earliness

Mori

Mendiburu

Dasgupta

et al. 2018

IEEE Trans. Neural Netw. Learning Syst.

147

View full text Add to dashboard Cite

The problem of early classification of time series appears naturally in contexts where the data, of temporal nature, is collected over time, and early class predictions are interesting or even required. The objective is to classify the incoming sequence as soon as possible, while maintaining suitable levels of accuracy in the predictions. Thus, we can say that the problem of early classification consists in optimizing two objectives simultaneously: accuracy and earliness. In this context, we present a method for early classification of time series based on combining a set of probabilistic classifiers together with a stopping rule. This stopping rule will act as a trigger and will tell us when to output a prediction or when to wait for more data, and it's main novelty lies in the fact that it is built by explicitly optimizing a cost function based on accuracy and earliness. We have selected a large set of benchmark datasets and 4 other state-ofthe-art early classification methods and we have evaluated and compared our framework obtaining superior results in terms of both earliness and accuracy.

show abstract

Section: Introductionmentioning

confidence: 99%

Early Classification of Time Series by Simultaneously Optimizing the Accuracy and Earliness

Mori

Mendiburu

Dasgupta

et al. 2018

IEEE Trans. Neural Netw. Learning Syst.

147

View full text Add to dashboard Cite

show abstract

“…Although in the study by [2,15] the importance of early classification on time series is identified and some encouraging results are shown, the study only solved early classification as a problem of classifying prefixes of sequences. [19] pointed out the challenge of early classification is to study the trade-off between the earliness and the accuracy of classification.…”

Section: Related Workmentioning

confidence: 91%

“…The ensemble classifier is capable of making predictions on incomplete data by viewing unavailable suffixes of sequences as missing features. Bregon et al (2006) applied a case-based reasoning method to classify time series to monitor the system failure in a simulated dynamic system. The KNN classifier is used to classify incomplete time series using various distances, such as Euclidean distance and Dynamic time warping (DTW) distance [17].…”

Section: Related Workmentioning

confidence: 99%

Early classification on time series

2011

View full text Add to dashboard Cite

In this paper, we formulate the problem of early classification of time series data, which is important in some time-sensitive applications such as health informatics. We introduce a novel concept of MPL (minimum prediction length) and develop ECTS (early classification on time series), an effective 1-nearest neighbor classification method. ECTS makes early predictions and at the same time retains the accuracy comparable with that of a 1NN classifier using the full-length time series. Our empirical study using benchmark time series data sets shows that ECTS works well on the real data sets where 1NN classification is effective.

show abstract

“…The classic case-based reasoning model [7], [8], divides a reasoning cycle into four stages: 1) Retrieving (Retrieve): aimed at finding and obtaining a previously saved case; the objective of this phase is to recover cases whose experience are potentially useful to resolve a new problem. This stage generally requires a combination of search and match techniques [9]; similarity measures are usually useful tools to find a case close to the search conducted.…”

Section: B Cbr Life Cyclementioning

confidence: 99%

Applying Case-Based Learning to Improve the Efficiency in the Web Service Compositions

Ingrid-Durley¹,

Guzmán-Luna²

2014

IJET

View full text Add to dashboard Cite

Abstract-Several planning techniques in artificial intelligence have been used to perform web service composition (semantic or not), but this process typically uses heuristics based planners combined with search techniques usually too expensive in time solution. In this article, we propose the use of case-based reasoning to reduce the computation times of composition; the model aims to infer from past experience a solution that would guide the selection process during a Web services composition. The proposed methodology also uses a classification defined by an algorithm of semantic similarity technique, in order to compare the new problem, with all previous problems. The previous problem with greatest similarity is accompanied by its corresponding solution and is used to specify which goals already achieved and what remain to achieve for the new problem. The result demonstrates greater efficiency, reducing the search space spending less time.Index Terms-Composition of semantic web services, planning in artificial intelligence and case-based learning, INDYGO. I. INTRODUCTIONFrom the perspective of service composition, there are few approaches which integrate learning models to improve various aspects of the task of Web service. A particular model which proposes the use of planning in Web service composition is the concurrent planning and execution model (INDYGO) [1], [2], which Universidad Nacional at Medellin Campus designed and implemented, and aims to produce semantic web service compositions in real time while handling incomplete data. This model, like most service composers, carries out solutions without taking into account experiences of previous solutions. Due to this, it was deemed valuable to propose within the INDYGO model the implementation of case-based learning techniques to enable the improvement of its efficiency in matters concerning the composition process itself that the planner is in charge of. It is important to highlight that even though the model adjusts to a previous development, this proposal may also be easily adapted and implemented in other approaches which under AI planning techniques enable semantic Web service composition since it is in charge of formulating its solution relying on commonly known elements implemented within this dominion.To amplify the proposal, this document is organized as follows: section two revises the reference framework related Jaime Guzmán-Luna is with the Director of SINTELWEB Research Group, Universidad Nacional de Colombia, Medellin Campus (e-mail: jaguzman@unal.edu.co).to the problem of service composition using AI planning techniques, specifically, centered on INDYGO. Section three, discusses and analyzes case-based learning concepts and proposes a representation of the solution starting from them; section four details the integration architecture of the CBR and INDYGO model and its functionality. Section five summarizes some results of the validation of the model, and section six presents conclusions and future work related to this proposal. II. A COMP...

show abstract

Early Fault Classification in Dynamic Systems Using Case-Based Reasoning

Cited by 27 publications

References 10 publications

Early Classification of Time Series by Simultaneously Optimizing the Accuracy and Earliness

Early Classification of Time Series by Simultaneously Optimizing the Accuracy and Earliness

Early classification on time series

Applying Case-Based Learning to Improve the Efficiency in the Web Service Compositions

Contact Info

Product

Resources

About