Quantum Neural Network Tuning and Performance Evaluation for a Breast Cancer Dataset

Model tuning with the optimization of pipeline configuration is a well-established practice for the development of machine learning models. However, this often entails an exhaustive search process, especially as the parameter space expands with increasing model complexity. In the emerging field of quantum machine learning (QML), there is limited literature on the effects of configuration parameters, especially quantum-specific ones, and their choices on model performance. To address this gap, here we present a study exploring the impacts of data scaling and configuration parameters in quantum neural network (QNN) development using beta regression. Our experiments with two benchmark datasets showed that a well-tuned QNN can achieve predictive performance comparable to its classical counterparts. Our findings also demonstrate useful reference points of QNN model tuning to support a more efficient parameter optimization process.