I developed SVM and QSVM on iris_data to see the potential advantage of QML over classical ML but I got QSVC with less accuracy and more computation time compare to SVC , can some please help how can I increase accuracy and optimize the computation time of QSVC. pasted below the code for reference #code for SVC svc=SVC(kernel='linear') start=time.time() svc.fit(X_train,y_train) time.time()-start. #svc with 0.008 computation time and with accuracy of 0.95
#code for QSVC service=QiskitRuntimeService() with Session(service=service,backend='simulator_statevector') as session: sampler1=Sampler(session=session) fidelity=ComputeUncompute(sampler=sampler1) fmap=ZZFeatureMap(feature_dimension=4,reps=2,entanglement='linear') kernel=FidelityQuantumKernel(feature_map=fmap,fidelity=fidelity) qsvc=QSVC(quantum_kernel=kernel) #time: 5 minutes and accuracy 0.73
How can I improve the performance of QSVC?