研究目的
To propose an efficient adaptive neuro-fuzzy classifier (NFC) for the identification and monitoring of electroencephalogram-based BCI for motor imagery (MI) tasks, integrating the Jaya optimization algorithm with adaptive neuro-fuzzy inference systems to enhance classification accuracy.
研究成果
The proposed Jaya-based NFC with SSCG and LH (JayaNFCSSCGLH) demonstrates superior performance in terms of classification accuracy and computation time compared to other NFCs under consideration. The LH-based feature selection method enhances classification accuracy by discarding irrelevant features, and the SSCG training algorithm reduces computational complexity. The classifier is suitable for real-time applications, as evidenced by its use in controlling LED switching based on classification results.
研究不足
The study focuses on two-class MI tasks and uses only two channels (C3 and C4) of EEG data, which may limit the generalizability of the results to more complex tasks or higher channel data.
1：Experimental Design and Method Selection:
The study integrates the Jaya optimization algorithm with adaptive neuro-fuzzy inference systems (ANFIS) for classifying two-class MI tasks. The linguistic hedge (LH) is used for proper elicitation and pruning of fuzzy rules, and the network is trained using scaled conjugate gradient (SCG) and speeding up SCG (SSCG) techniques.
2：Sample Selection and Data Sources:
EEG signals are recorded from nine right-handed subjects using bipolar recording from the motor cortex area (C3, C4, and Cz). The signals are filtered and features are extracted for μ (7-14 Hz) and ? (17-26 Hz) bands.
3：List of Experimental Equipment and Materials:
MP 100 from Biopac Inc. is used for signal recording, and Acknowledge 3.9 is used as the interface development environment (IDE).
4：9 is used as the interface development environment (IDE).
Experimental Procedures and Operational Workflow:
4. Experimental Procedures and Operational Workflow: The Jaya-based k-means algorithm is applied to divide the feature set into two mutually exclusive clusters and fire the fuzzy rule. The performance of the proposed classifier is compared with four different NFCs.
5：Data Analysis Methods:
The classification accuracy and computation time are analyzed, and the Friedman test is used to validate the effectiveness of the proposed technique.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
