Hybrid Quantum-Classical k-NN Model for Improved Accuracy
Defence Science Journal · 29.04.2026 · 76 (3) · 379-389
DOI: 10.14429/dsj.21511 ↗Kurz gefasst
Fachartikel in der Zeitschrift Defence Science Journal, online erschienen am 29.04.2026. Verfasst von Prashant Verma und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 2.5).
Abstract (Originalsprache)
k-Nearest Neighbours (k-NN) is a widely used non-parametric, instance-based classification algorithm; however, its traditional implementation suffers from high computational cost, scalability limitations, and performance degradation in high-dimensional spaces. To address these challenges, this paper proposes a hybrid quantum-classical k-Nearest Neighbours (Qk-NN) model that integrates quantum feature encoding and quantum-assisted feature transformation with classical k-NN classification. The proposed approach is evaluated on the Pima Indians Diabetes Dataset using feature subsets of 5, 6, and 7 dimensions, selected due to quantum circuit scalability constraints. Quantum feature transformations are implemented through parameterized quantum circuits simulated on classical hardware in an ideal noiseless environment. Model performance is assessed using accuracy and log loss with a fixed Neighbourhood size of k = 5. Experimental results show that the Hybrid Qk-NN is superior to the classical k-NN in each feature subset. In particular, with five features, the Hybrid Qk-NN registers an accuracy of 77. 62 % while classical k-NN only achieves 71. 84 %, meanwhile, the loss values are quite lower. Quantum-assisted feature mapping capable of increasing the performance of classification is suggested by these results, and hybrid quantum-classical models are shown to be a viable bridge to the quantum machine learning world in the future.
Zitieren
APA
Verma, P., & Vishwakarma, D. K. (2026). Hybrid Quantum-Classical k-NN Model for Improved Accuracy. Defence Science Journal, 76(3), 379-389. https://doi.org/10.14429/dsj.21511
Vancouver
Verma P, Vishwakarma DK. Hybrid Quantum-Classical k-NN Model for Improved Accuracy. Defence Science Journal. 2026;76(3):379-389. doi:10.14429/dsj.21511
BibTeX
@article{Verma2026,
author = {Verma, Prashant and Vishwakarma, Dinesh Kumar},
title = {Hybrid Quantum-Classical k-NN Model for Improved Accuracy},
journal = {Defence Science Journal},
year = {2026},
volume = {76},
number = {3},
pages = {379--389},
doi = {10.14429/dsj.21511}
}
EUVIMED