Results

The results of this module highlight a recurring theme in contemporary Quantum Machine Learning research: classical models remain formidable baselines. The MLP (93% accuracy), despite its simplicity, benefits from access to all 8 features and highly optimized gradient computation, giving it a natural advantage on structured tabular data. 

The VQC (66% accuracy), constrained to 4 qubits and sequential circuit evaluation, trades raw performance for a fundamentally different computational approach, one that may scale more favorably as quantum hardware matures. 

And finally, the hybrid model (with 96% accuracy) attempts to get the best of both worlds, using classical layers to handle the dimensionality mismatch and letting the quantum layer focus on feature transformation, though whether the quantum component genuinely adds expressibility or simply acts as a bottleneck depends heavily on the dataset and circuit design.