EEG Data Fusion for Improving Accuracy of Binary Classification.

Area of Science:

• Biomedical Engineering
• Data Science
• Signal Processing

Background:

• Accurate classification of multiple medical data is crucial for diagnosis and treatment.
• Existing algorithms for electroencephalogram (EEG) data processing face challenges in real-time applications.
• Data fusion techniques are essential for integrating information from various sources.

Purpose of the Study:

• To propose a novel methodology for the classification of multiple medical data, specifically focusing on EEG signals.
• To evaluate the effectiveness of different variations of the Dempster-Shafer technique for data fusion in this context.
• To assess the potential for real-time data classification using the proposed approach.

Main Methods:

• A seven-stage methodology was developed for EEG data processing.
• The Dempster-Shafer technique and its variations were employed as the core data fusion instrument.
• Performance was evaluated based on classification accuracy.

Main Results:

• The proposed methodology achieved classification accuracy comparable to established algorithms.
• The data fusion approach using Dempster-Shafer variations proved effective for EEG data.
• The system demonstrated potential for real-time data classification.

Conclusions:

• The developed seven-stage methodology offers a viable approach for multiple medical data classification.
• The Dempster-Shafer technique provides a robust framework for data fusion in EEG analysis.
• This work lays a foundation for future advancements in real-time medical data classification systems.