Pairwise classification using quantum support vector machine with Kronecker kernel

Abstract We investigated the potential application of quantum computing using Kronecker kernel to pairwise classification and have devised a way apply Harrow-Hassidim-Lloyd (HHL)-based support vector machine algorithm. Pairwise can be used predict relationships among data is for problems such as link prediction chemical interaction prediction. However, in kernel, it very costly calculate product matrices when there large amount data. found that represented more efficiently time space than classical computing. also classifier effectively trained by applying HHL-based algorithm matrix. In an experiment comparing with run on simulator, misclassification rate latter was almost same former problem some cases. This indicates achieve accuracy equivalent scalably. finding paves learning predicting large-scale