Takagi--Sugeno--Kang Fuzzy Classifiers for a Special Class of Time-Varying Systems

Mikut, Ralf; Burmeister, Ole; Gröll, Lutz; Reischl, Markus

doi:10.1109/tfuzz.2008.917291

Cited by 10 publications

(7 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Table 7). An error of 14.2 percent is achieved with TSK fuzzy classifiers in [33], another approach cited in [33] leads to 15.8 percent misclassified time series. The best classification results given in [55] are at about 10.6 percent.…”

Section: Ecg Time Seriesmentioning

confidence: 98%

“…Fuzzy techniques inherently support the description of a classifier with a set of comprehensible rules. Examples in the field of time series classification can be found in [32], [33], [34]. Often, final pruning steps are needed to reduce the number of generated rules and to eliminate conflicting rules.…”

Section: Related Workmentioning

confidence: 99%

Section: Cbf Time Seriesmentioning

confidence: 99%

“…An approach based on Takagi-Sugeno-Kang (TSK) fuzzy classifiers is proposed in [33] with 0.23 percent classification error. Some other approaches that are referenced in [33] yield errors between 0 percent and 0.75 percent. A new distance measure for time series (the triangle distance) is suggested in [52] and the best classification error is 0 percent using a 1-NN (nearest neighbor) classifier.…”

Section: Cbf Time Seriesmentioning

confidence: 99%

See 3 more Smart Citations

SwiftRule: Mining Comprehensible Classification Rules for Time Series Analysis

Fisch

Gruber

Sick

2011

IEEE Trans. Knowl. Data Eng.

View full text Add to dashboard Cite

In this article, we provide a new technique for temporal data mining which is based on classification rules that can easily be understood by human domain experts. Basically, time series are decomposed into short segments, and short-term trends of the time series within the segments (e.g., average, slope, and curvature) are described by means of polynomial models. Then, the classifiers assess short sequences of trends in subsequent segments with their rule premises. The conclusions gradually assign an input to a class. As the classifier is a generative model of the processes from which the time series are assumed to originate, anomalies can be detected, too. Segmentation and piecewise polynomial modeling are done extremely fast in only one pass over the time series. Thus, the approach is applicable to problems with harsh timing constraints. We lay the theoretical foundations for this classifier, including a new distance measure for time series and a new technique to construct a dynamic classifier from a static one, and demonstrate its properties by means of various benchmark time series, for example, Lorenz attractor time series, energy consumption in a building, or ECG data.

show abstract

Section: Ecg Time Seriesmentioning

confidence: 98%

Section: Related Workmentioning

confidence: 99%

Section: Cbf Time Seriesmentioning

confidence: 99%

Section: Cbf Time Seriesmentioning

confidence: 99%

See 2 more Smart Citations

SwiftRule: Mining Comprehensible Classification Rules for Time Series Analysis

Fisch

Gruber

Sick

2011

IEEE Trans. Knowl. Data Eng.

View full text Add to dashboard Cite

show abstract

“…B. für die Klasse 1 folgende Heuristik angewendet: Außerdem können durch geeignete Vorverarbeitungsschritte die Anzahl der nötigen Klassifikatoren deutlich reduziert werden [32].…”

unclassified

Zeitvariante Klassifikatoren zur Steuerung von Brain Machine Interfaces und Neuroprothesen (Time-variant Classifiers to Control Brain Machine Interfaces and Neuroprostheses)

Burmeister¹,

Reischl²,

Gröll³

et al. 2006

At - Automatisierungstechnik

Self Cite

View full text Add to dashboard Cite

Bretthauer zum 60. Geburtstag gewidmet Der Beitrag analysiert Ursachen für zeitvariante Klassifikationsprobleme am Beispiel von Brain Machine Interfaces, die zur Steuerung technischer Systeme mit Hirnsignalen dienen. Ausgehend von einer systematischen Kategorisierung der Integration zeitvarianter Informationen in Klassifikatoren werden unterschiedliche Konzepte vorgeschlagen und anhand dreier Benchmarkdatensätze der Brain Computer Interface Competition III (2005) erfolgreich erprobt.This paper analyzes reasons for time-variant classification problems considering brain machine interfaces (i. e., interfaces to control devices by brain signals) as an example. Starting from a categorization for integrating time-variant information in classifiers, different concepts of time-variant classifiers are proposed and successfully applied using three benchmark datasets from the Brain Computer Interface Competition III (2005).Schlagwörter: Zeitvariante Klassifikatoren, zeitvariante Probleme, Brain Machine Interface, Neuroprothesen

show abstract