Efficient Epileptic Seizure Detection using K-Nearest Neighbors and Wavelet-based Feature Extraction from EEG Signals

EEG recording detection of epileptic seizures is essential for helping doctors diagnose and treat patients in a timely manner. This paper presents a useful method that combines the K-Nearest Neighbors (KNN) classification method with wavelet-based feature extraction. To identify important patterns at various frequency levels, EEG signals are first processed using the Discrete Wavelet Transform (DWT). In order to differentiate seizure activity, statistical features are extracted from the converted signals. The KNN algorithm is then used to classify these features, determining the most likely category based on proximity to known samples. All important metrics showed that the system performed well: 94.2% accuracy, 93.1% precision, 92.7% recall, 92.9% F1-score, and an AUC-ROC of 0.96. The results indicate that this approach is not only efficient but also suitable for real-world implementation in clinical or portable monitoring systems.