Time series for blind biosignal classification model.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Machine Learning

Background:

• Biosignals (ECG, EEG, EMG) are crucial for noninvasive diagnostics.
• Automated biosignal classification aids disease diagnosis.
• Current methods require prior knowledge of biosignal type.

Purpose of the Study:

• To propose a blind biosignal classification model (B(2)SC Model).
• To automatically identify the source biosignal type without prior information.
• To enhance diagnostic decision-making by enabling classification of unknown biosignal types.

Main Methods:

• Utilized time series algorithms for model construction.
• Employed dynamic time warping (DTW) with clustering.
• Discovered similarities between biosignals to classify disease.

Main Results:

• Demonstrated the effectiveness of the B(2)SC Model.
• Showcased the scalability of the proposed approach.
• Successfully classified biosignals without prior type identification.

Conclusions:

• The B(2)SC Model offers a novel solution for classifying unknown biosignal types.
• This method can advance automated disease diagnosis by handling diverse biosignals.
• The approach is effective and scalable for real-world applications.