Quantum support vector machine without iteration

Quantum algorithms can enhance machine learning in different aspects. The quantum support vector was proposed to improve the performance, which Swap Test plays a crucial role realizing classification. However, as is destructive, must be repeated preparing qubits and manipulating operations. This paper proposes based on amplitude estimation (AE-QSVM) gets rid of constraint repetitive process saves resources. At first, generalized introduced initial state arbitrary instead being |0〉. Then, AE-QSVM trained by singular value decomposition query sample classified estimation. In AE-QSVM, high accuracy achieved adding auxiliary repeating algorithm. time space complexity are reduced compared with other algorithms. Finally, we ran experiments IBM's computer experimental results demonstrate that classification 95% probability success only uses 12 qubits.