socialprescribing/quantum-ltc-qltcmlm

QLTCMLM: 4-Qubit Frobenius Quantum Long-Term Conditions Model (NHS NCL LCS v6 Research Prototype)

Research-only multi-task 4-qubit Frobenius data-reuploading Variational Quantum Classifier (QLTCMLM-VQC) and Fidelity Quantum Kernel (QLTCMLM-QKernel) trained on Modal Cloud for synthetic NHS North Central London ICB Long Term Conditions (LTC) risk re-stratification and Year-of-Care bundle recommendation (phi_present = False, 0 physical QPU shots claimed; all NHS DCB0129 clinical hazards remain OPEN).

Benchmark Results (N = 2,400 Non-Linear Cubic-Interaction Cohort, 25% Held-Out Test)

Model	Risk Tier Accuracy	Care Bundle Accuracy
QLTCMLM-VQC (4-Qubit, 3-Layer Frobenius Data-Reuploading) (200 total params: 72 quantum + 128 readout)	0.5417	0.7550
**QLTCMLM-QKernel (4-Qubit Fidelity Gram Matrix	<psi(x)	psi(x')>
LogisticRegression (Linear Hyperplane)	0.3050	0.6850
RBF-SVM (Classical Kernel C=4.0)	0.6767	0.8500
MLPClassifier (2-Layer 32x16 ReLU)	0.6350	0.8017
HistGradientBoosting (GBDT Baseline)	0.5367	0.8183