A window-based time series feature extraction method

Katircioglu-ztrk, Deniz; Gvenir, H. Altay; Ravens, Ursula; Baykal, Nazife

doi:10.1016/j.compbiomed.2017.08.011

Cited by 11 publications

(4 citation statements)

References 58 publications

(77 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For obtaining the appropriate trend of features, window-wise feature extraction was needed. 26,27 After selecting three selected endodontic files, good, partially worn, and severely worn files, a window was employed in each time-domain signal, as shown in Figure 5. Features were extracted according to the window-wise feature extraction process.…”

Section: Methodsmentioning

confidence: 99%

“…Various techniques can define the health index. A few of them include principal component analysis (PCA)/Classical PCA, 28,29 kernel PCA, 28 diffusion mapping (DM), 28 window wise feature extraction, 26 maximum variance unfolding (MVU), 28 self-organizing mapping (SOM) network, 30 and isometric feature mapping reduction technique (ISOMAP). 31 In this work, the health index was determined using the widow-wise method.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Health prediction of reciprocating endodontic instrument based on the machine learning and exponential degradation models

Thakur,

Kankar,

Parey

et al. 2023

Proc Inst Mech Eng H

View full text Add to dashboard Cite

This study proposes an intelligent health prediction and fault prognosis of the endodontic file during the root canal treatment. Root canal treatment is the procedure of disinfecting the infected pulp through the canal with the help of an endodontic instrument. Force signals are acquired with the help of a dynamometer during the canal preparation, and statistical features are extracted. The extracted features are selected through the window-wise feature extraction process. Characteristic features for endodontic file prognostics include time-domain features of the signals are evaluated. The extracted feature has inappropriate information, that is, noise between the signals; hence the smoothing of the feature is required at this stage to observe a trend in the signals. Based on the smoothing feature and post-processing of the feature, defined the health index to calculate the health condition of the endodontic instruments. A machine learning algorithm and exponential degradation model are used to predict the health of the endodontic instrument during the root canal treatment. This model is used to forecast the degradation of the endodontic file so that actions can be taken before actual failures happen. The proposed methodology can analyze the failures and micro-crack initiation of the endodontic instruments. Endodontics practitioners can use the machine learning models as well as an exponential model for estimating the health condition of the endodontic instrument. This study may help the clinician to progress the efficiency of the root canal treatment and the competence of the endodontic instruments.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Health prediction of reciprocating endodontic instrument based on the machine learning and exponential degradation models

Thakur,

Kankar,

Parey

et al. 2023

Proc Inst Mech Eng H

View full text Add to dashboard Cite

show abstract

“…Oztürk et al (2017) developed a feature extraction technique which is based on similarity score. The technique called 'WTC' or Window based Time Series Feature Extraction, and applied on human cardiomyocytes and ECG dataset [6].…”

Section: Literature Reviewmentioning

confidence: 99%

Feature Extraction and Clustering of Twenty Three IIPs of India

Das¹,

Mukherjee²,

Roy³

et al. 2019

IJRAT

View full text Add to dashboard Cite

The study analyzes the 'Index of Industrial Production' of twenty three (23) different items of India by extracting the trend, spike, curvature, linearity and entropy features of each of the twenty three (23) indexes. The authors also cluster this time series data on the basis of five (5) extracted features. The largest clusters of indexes were exhibited by both the linearity and spike features. For each of them i.e. linearity and spike features, the larger cluster contains nineteen ( 19) indexes. Outlier clusters were also identified in the cases of trend, linearity, spike and curvature features. The cluster analysis of the indexes together on spike and linearity gave us a single cluster with twenty (20) indexes while three (3) indexes were marked as outlier cluster. Index Terms-Feature extraction; Clustering; Time Series; IIP.

show abstract

“…Many methods for automatic classification of ECGs have been proposed. The type of ECG beat can be distinguished by the time-domain [4], wavelet transform [5], genetic algorithm [6], support vector machine (SVM) [7], Bayesian [8], or other methods. Although the above classification methods achieve high accuracy on experimental datasets, their performance is highly dependent on the extraction characteristics of fixed or manual design methods.…”

Section: Introductionmentioning

confidence: 99%

An Effective LSTM Recurrent Network to Detect Arrhythmia on Imbalanced ECG Dataset

Gao

Zhang

et al. 2019

Journal of Healthcare Engineering

115

View full text Add to dashboard Cite

To reduce the high mortality rate from cardiovascular disease (CVD), the electrocardiogram (ECG) beat plays a significant role in computer-aided arrhythmia diagnosis systems. However, the complex variations and imbalance of ECG beats make this a challenging issue. Since ECG beat data exist in heavily imbalanced category, an effective long short-term memory (LSTM) recurrence network model with focal loss (FL) is proposed. For this purpose, the LSTM network can disentangle the timing features in complex ECG signals, while the FL is used to resolve the category imbalance by downweighting easily identified normal ECG examples. The advantages of the proposed network have been verified in the MIT-BIH arrhythmia database. Experimental results show that the LSTM network with FL achieved a reliable solution to the problem of imbalanced datasets in ECG beat classification and was not sensitive to quality of ECG signals. The proposed method can be deployed in telemedicine scenarios to assist cardiologists into more accurately and objectively diagnosing ECG signals.

show abstract

A window-based time series feature extraction method

Cited by 11 publications

References 58 publications

Health prediction of reciprocating endodontic instrument based on the machine learning and exponential degradation models

Health prediction of reciprocating endodontic instrument based on the machine learning and exponential degradation models

Feature Extraction and Clustering of Twenty Three IIPs of India

An Effective LSTM Recurrent Network to Detect Arrhythmia on Imbalanced ECG Dataset

Contact Info

Product

Resources

About