Geometric Analysis for QSVM Application using Kullback-Leibler Divergence

In supervised quantum machine learning, mapping data to quantum states is crucial. We propose a novel method to analyze Quantum Feature Maps by examining quantum states from data mapping, considering each qubit's projections onto Pauli observable values. Using Kullback-Leibler divergence, we analyze quantum states for binary data in Quantum Support Vector Machines. Our findings show a significant relationship between Kullback-Leibler divergence, data separability, and model accuracy, highlighting our approach's effectiveness.