 This paper proposes a novel hybrid system combining an unsupervised learning, UL, module and a supervised learning, SL, module to detect epileptic seizures from electroencephalography, EEG, signals. The UL module reduces the manual effort required for data labeling, while the SL module improves the accuracy of the detection process. The proposed system was tested on the CHBMIT dataset and achieved a mean accuracy of 92.62 mean sensitivity of 95.55% and a mean specificity of 92.57%. This is the first study to combine both UL and SL modules for epileptic seizure detection, demonstrating its potential for reducing the manual effort required for data labeling and improving the accuracy of the detection process. This article was authored by Yao Gua, Sinyu Jiang, Lin Kai Tao and others.