Quantum machine learning for support vector machine classification

Quantum machine learning aims to execute machine learning algorithms in quantum computers by utilizing powerful laws like superposition and entanglement for solving problems more efficiently. Support vector machine (SVM) is proved to be one of the most efficient classification machine learning algorithms in today’s world. Since in classical systems, as datasets become complex or mixed up, the SVM kernel approach tends to slow and might fail. Hence our research is focused to examine the execution speed and accuracy of quantum support vector machines classification compared to classical SVM classification by proper quantum feature mapping selection. As the size of the dataset becomes complex, a proper feature map has to be selected to outperform or equally perform the classification. Hence the paper focuses on the selection of the best feature map for some benchmark datasets. Additionally experimental results show that the processing time of the algorithm is considerably reduced concerning classical machine learning. For evaluation of quantum computation over the classical computer, Quantum labs from the IBMQ quantum computer cloud have been used.