Quantum-enhanced Hybrid Signal Reconstruction Framework for Tuberculosis Epidemiological Time Series
Year: 2026
Type: article
Abstract: Tuberculosis (TB) epidemiological time series often show complex, non-stationary, and multi-scale dynamics driven by exogenous factors, complicating the separation of informative signals from stochastic fluctuations.This study proposes a Quantum-Enhanced Hybrid (QH) framework for adaptive denoising that integrates multiresolution wavelet decomposition, Quantum Mutual Information (QMI), Bayesian Optimization (BO), and modified Quantum-Inspired Optimization Algorithm-Quantum-behaved Particle Swarm Optimization (mQIOA-QPSO).QMI guides threshold tuning across wavelet scales, while BO explores global configurations and mQIOA-QPSO refines them locally.Using five-fold Time Series Cross-Validation (TSCV) and observed TB incidence as an empirical reconstruction reference, QH achieved the best numerical profile among Empirical Mode Decomposition (EMD), Universal Wavelet, and Singular Value Decomposition (SVD), with RMSE=52.56,MAE=33.61,R²=0.940,SNR=12.20,PSNR=25.21,Corr=0.948, and SpecCorr=0.992.QH reduced RMSE by 29.0%, 42.2%, and 65.1% relative to the baselines.Friedman testing indicated statistically significant overall differences, whereas Nemenyi posthoc testing showed pairwise significance mainly between QH and SVD.
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7204668780
FWCI:
Cites:
Cited by:
Related to: