Unsupervised Classification of Epileptic EEG Signals with Multi Scale K-Means Algorithm

Zhu, Guohun; Li, Yan; Wen, Peng; Wang, Shuaifang; Zhong, Ning

doi:10.1007/978-3-319-02753-1_16

Cited by 13 publications

(8 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The highest overall classification accuracies are highlighted in bold font. The best performance was obtained from the OA technique combined with the AR estimation method as well as SVM classifier, compared with the results obtained by the authors of [16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31]. Even though the extracted features by using the OA combined with the mentioned spectral methods were scored the second highest accuracy with a 99.9%, it is considered acceptable in the research of epileptic seizures detection.…”

Section: Resultsmentioning

confidence: 73%

“…Hassan and Subasi [21] obtained 100% accuracy for only A versus E sets by using complete ensemble empirical mode decomposition and adaptive noise with a linear programming boosting compared with our method, which used all datasets and yielded 100% accuracy. Although the achieved results by the authors of [22, 27, 28] were 100% accuracies, they used part of datasets that was A group versus E group. The detection performance of the proposed technique was also higher than those by the authors of [24–26, 30, 31].…”

Section: Resultsmentioning

confidence: 99%

“…A versus E 99.82 A, B, C, D versus E 98.72 A versus E 99.80 A, B, C, D versus E 98.65 Zhu et al [27] • sample entropy + multi-scale K-means…”

Section: Acknowledgmentmentioning

confidence: 99%

“…From our team's work, Zhu et al . [27] proposed a multi‐scale K ‐means approach to detect the epileptic seizures from EEG signals. In order to find these seizures, the researchers extracted six key features by using sample entropy method and yielded 100, 99, and 99.1% classification accuracies for three cases (A versus E, B versus E, and AB versus CDE), respectively.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Epileptic EEG signal classification using optimum allocation based power spectral density estimation

Ghayab

Siuly

et al. 2018

IET signal process.

Self Cite

View full text Add to dashboard Cite

This study proposes a novel approach blending optimum allocation (OA) technique and spectral density estimation to analyse and classify epileptic electroencephalogram (EEG) signals. This study employs the OA to determine representative sample points from the original EEG data and then applies periodogram (PD), autoregressive (AR), and the mixture of PD and AR to extract the discriminative features from each OA sample group. The obtained feature sets are evaluated by three popular machine learning methods: support vector machine (SVM), quadratic discriminant analysis (QDA), and k-nearest neighbour (k-NN). Several output coding approaches of the SVM classifier are tested for selecting the best feature sets. This scheme was implemented on a benchmark epileptic EEG database for evaluation and also compared with existing methods. The experimental results show that the OA_AR feature set yields better performances by the SVM with an overall accuracy of 100%, and outperforms the state-of-the-art works with a 14.1% improvement. Thus, the findings of this study prove that the proposed OA-based AR scheme has significant potential to extract features from EEG signals. The proposed method will assist experts to automatically analyse a large volume of EEG data and benefit epilepsy research.

show abstract

Section: Resultsmentioning

confidence: 73%

Section: Resultsmentioning

confidence: 99%

“…A versus E 99.82 A, B, C, D versus E 98.72 A versus E 99.80 A, B, C, D versus E 98.65 Zhu et al [27] • sample entropy + multi-scale K-means…”

Section: Acknowledgmentmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Epileptic EEG signal classification using optimum allocation based power spectral density estimation

Ghayab

Siuly

et al. 2018

IET signal process.

Self Cite

View full text Add to dashboard Cite

show abstract

“…K-means is the commonly used method in practice [13]. In the study [14], Support Vector Machine and K-means with Multi-Scale K-means (MSK-means) were compared. Two classes were detected (epileptic and non-epileptic) with the best results for MSK-means.…”

Section: Introductionmentioning

confidence: 99%

Automatic Eeg Classification Using Density Based Algorithms Dbscan and Denclue

2019

View full text Add to dashboard Cite

Electroencephalograph (EEG) is a commonly used method in neurological practice. Automatic classifiers (algorithms) highlight signal sections with interesting activity and assist an expert with record scoring. Algorithm K-means is one of the most commonly used methods for EEG inspection. In this paper, we propose/apply a method based on density-oriented algorithms DBSCAN and DENCLUE. DBSCAN and DENCLUE separate the nested clusters against K-means. All three algorithms were validated on a testing dataset and after that adapted for a real EEG records classification. 24 dimensions EEG feature space were classified into 5 classes (physiological, epileptic, EOG, electrode, and EMG artefact). Modified DBSCAN and DENCLUE create more than two homogeneous classes of the epileptic EEG data. The results offer an opportunity for the EEG scoring in clinical practice. The big advantage of the proposed algorithms is the high homogeneity of the epileptic class.

show abstract