Selected work

Representative Papers

Q-Sat AI: Machine Learning-Based Decision Support for Data Saturation in Qualitative Studies

Nov 02, 2025

Qualitative research often relies on subjective judgments of data saturation to determine sample size, leading to methodological inconsistency and compromised rigor. To address this, we propose the first machine learning–based decision support model for sample size determination, pioneering the application of ensemble learning methods—including XGBoost and Random Forest—to quantitatively model the data saturation process. The model integrates ten key study design parameters, undergoes rigorous preprocessing and outlier removal, and achieves an R² of 0.85, effectively capturing nonlinear relationships in sampling dynamics. Feature importance analysis empirically validates foundational theoretical assumptions—such as the influence of study type and informational power—advancing standardization in qualitative methodology. The model has been implemented as an open-source web tool for researchers and reviewers, substantially enhancing transparency, reproducibility, and methodological rigor in sample size justification.

0 citationsRead paper

Quantum Neural Networks for Wind Energy Forecasting: A Comparative Study of Performance and Scalability with Classical Models

Jun 28, 2025

Accurate short-term wind power forecasting is critical for smart grid stability, yet quantum machine learning (QML) applications in renewable energy modeling remain underexplored. Method: This study systematically evaluates the practicality and scalability of quantum neural networks (QNNs) for wind power forecasting, introducing the first empirical benchmark specifically for this domain. We comparatively assess multiple Z-feature-mapped QNN architectures—incorporating diverse variational ansätze—within a unified experimental framework, employing cross-validation and independent test sets. Contribution/Results: QNNs achieve prediction accuracy comparable to state-of-the-art classical models (e.g., LSTM, XGBoost), with marginal superiority in certain data regimes. Crucially, we quantify the exponential growth in classical simulation time with circuit depth and qubit count. This work establishes the first domain-specific QML benchmark for wind forecasting, delineating current hardware-imposed limitations on QNN deployment and identifying concrete optimization pathways toward practical quantum advantage in energy systems.

0 citationsRead paper

Quantum-Enhanced Classification of Brain Tumors Using DNA Microarray Gene Expression Profiles

May 04, 2025

High-dimensional DNA microarray gene expression data (54,676 features) pose significant challenges for multi-class brain tumor classification, including the curse of dimensionality, difficulty in modeling complex nonlinear patterns, and low computational efficiency. Method: We propose a Deep Variational Quantum Classifier (Deep VQC), the first quantum machine learning framework tailored for brain tumor diagnosis. It integrates quantum superposition and entanglement into a hybrid quantum-classical training architecture and introduces a biologically informed feature embedding and preprocessing pipeline specifically designed for high-dimensional genomic data. Contribution/Results: Evaluated on a five-class task—comprising four brain tumor subtypes and healthy controls—Deep VQC achieves accuracy comparable to or exceeding state-of-the-art classical models (e.g., SVM, XGBoost, DNN). It demonstrates markedly improved generalization and inference efficiency in high-dimensional, small-sample regimes. This work establishes the feasibility and translational potential of variational quantum machine learning in precision medical diagnostics.

0 citationsRead paper

Recent publications

Latest Papers

Q-Sat AI: Machine Learning-Based Decision Support for Data Saturation in Qualitative Studies

Nov 02, 2025

Qualitative research often relies on subjective judgments of data saturation to determine sample size, leading to methodological inconsistency and compromised rigor. To address this, we propose the first machine learning–based decision support model for sample size determination, pioneering the application of ensemble learning methods—including XGBoost and Random Forest—to quantitatively model the data saturation process. The model integrates ten key study design parameters, undergoes rigorous preprocessing and outlier removal, and achieves an R² of 0.85, effectively capturing nonlinear relationships in sampling dynamics. Feature importance analysis empirically validates foundational theoretical assumptions—such as the influence of study type and informational power—advancing standardization in qualitative methodology. The model has been implemented as an open-source web tool for researchers and reviewers, substantially enhancing transparency, reproducibility, and methodological rigor in sample size justification.

0 citationsRead paper

Quantum Neural Networks for Wind Energy Forecasting: A Comparative Study of Performance and Scalability with Classical Models

Jun 28, 2025

Accurate short-term wind power forecasting is critical for smart grid stability, yet quantum machine learning (QML) applications in renewable energy modeling remain underexplored. Method: This study systematically evaluates the practicality and scalability of quantum neural networks (QNNs) for wind power forecasting, introducing the first empirical benchmark specifically for this domain. We comparatively assess multiple Z-feature-mapped QNN architectures—incorporating diverse variational ansätze—within a unified experimental framework, employing cross-validation and independent test sets. Contribution/Results: QNNs achieve prediction accuracy comparable to state-of-the-art classical models (e.g., LSTM, XGBoost), with marginal superiority in certain data regimes. Crucially, we quantify the exponential growth in classical simulation time with circuit depth and qubit count. This work establishes the first domain-specific QML benchmark for wind forecasting, delineating current hardware-imposed limitations on QNN deployment and identifying concrete optimization pathways toward practical quantum advantage in energy systems.

0 citationsRead paper

Quantum-Enhanced Classification of Brain Tumors Using DNA Microarray Gene Expression Profiles

May 04, 2025

High-dimensional DNA microarray gene expression data (54,676 features) pose significant challenges for multi-class brain tumor classification, including the curse of dimensionality, difficulty in modeling complex nonlinear patterns, and low computational efficiency. Method: We propose a Deep Variational Quantum Classifier (Deep VQC), the first quantum machine learning framework tailored for brain tumor diagnosis. It integrates quantum superposition and entanglement into a hybrid quantum-classical training architecture and introduces a biologically informed feature embedding and preprocessing pipeline specifically designed for high-dimensional genomic data. Contribution/Results: Evaluated on a five-class task—comprising four brain tumor subtypes and healthy controls—Deep VQC achieves accuracy comparable to or exceeding state-of-the-art classical models (e.g., SVM, XGBoost, DNN). It demonstrates markedly improved generalization and inference efficiency in high-dimensional, small-sample regimes. This work establishes the feasibility and translational potential of variational quantum machine learning in precision medical diagnostics.

0 citationsRead paper