Hyper parameter optimization with quantum programming

Hyperparameter optimization, a crucial problem in the field of machine learning, directly affects model performances. It has been observed in studies that hyperparameter optimization positively influences the success of machine learning algorithms and provides optimal time costs in models with complex search spaces. In this study, this problem has been approached using classical, optimization, and quantum optimization methods. The use of quantum computing in machine learning methods, a rapidly advancing topic in recent years, has attracted significant interest. This latest technological approach, situated at the intersection of quantum physics and machine learning, leverages the principles of quantum mechanics to process information at scales beyond the capabilities of classical computers. The hyperparameter optimization methods included in this study are Grid Search, Random Search, and Bayesian Optimization; Genetic Algorithm and Particle Swarm Optimization; Quantum Genetic Algorithm and Quantum Particle Swarm Optimization. Experiments were conducted on the hyperparameters of the classic machine learning algorithm Support Vector Machine and the hybrid model with quantum features, the Quantum Support Vector Machine. In light of the obtained results, while the Grid Search method provided the best performance, it was observed that the optimization methods also demonstrated quite close results to this performance.