A Hybrid Regression and Parameterized Quantum Learning Model for EEG ‐Based Seizure Classification in IoT ‐Enabled Smart Healthcare

The integration of Internet of Things technologies into smart healthcare systems has revolutionized the continuous monitoring of neurological conditions through wearable electroencephalogram devices. However, electroencephalogram signals are noisy, nonstationary, and highly dimensional, which limits the efficiency of conventional models. To address these challenges, this paper proposes a novel hybrid model termed Theil‐Sen Correlated Least Squares Regression and Parameterized Quantum Deep Learning model for electroencephalogram seizure classification. The proposed model integrates three major components: (i) Flexible Wavelet Packet Decomposition for denoising and time‐frequency feature extraction, (ii) Modified Theil‐Sen Pearson Correlation Least Square Regression method for feature selection, and (iii) a Parameterized Quantum Circuit Long Short‐Term Memory classifier for the precise classification of temporal signals. The quantum component is implemented using a classical state vector simulator. Hence, the proposed model improves the classification accuracy, reduces computational overhead, and improves real‐time responsiveness. The experimental evaluations performed on the three publicly available benchmark electroencephalogram datasets exhibited efficient performance. The Theil‐Sen Correlated Least Squares Regression and Parameterized Quantum Deep Learning model achieves an accuracy of 98.89%, precision of 97.92%, recall of 98.11%, and 25.23 ms inference time. Thus, the experimental results demonstrate that the proposed model achieves superior classification performances when compared with existing models.