Published Paper Details:

The Friedman-based fuzzy rough quick reduct algorithm(FMFQR) is a feature selection method designed to enhance the efficiency and accuracy of machine learning models. By integrating Friedman's statistical test with fuzzy rough set theory, the algorithm identifies and eliminates irrelevant or redundant features while preserving essential data characteristics. This approach combines the strengths of fuzzy rough sets in handling uncertainty with Friedman's test to prioritize features based on their significance. The quick reduct process ensures computational efficiency, making it suitable for high-dimensional datasets. In this paper, First, Statistical feature extraction method is applied for raw EEG data. Then the extracted features are normalized by Z-Score normalization approach, Further, the normalized features are selected by Friedman-based fuzzy rough quick reduct algorithm. Finally, the classification task is done using three classification algorithms such as Fuzzy-Rough Nearest Neighbor (FRNN), Fuzzy Logic-based Neural Networks (FNN), Fuzzy Decision Tree (FDT). The experimental works are discussed and analyzed. Results indicate that the proposed approach significantly improves a FRNN based classification method in terms of classification accuracy. The proposed method improves the accuracies when compared to other state of the methods.