Solving Feature Selection Problem by Quantum Optimization Algorithm

Tóm tắt báo cáo

This study is to propose a method for feature selection with Quadratic Unconstrained Binary Optimization (QUBO) formulation by applying Conditional Value at Risk (CVaR) hybrid with Quantum Approximate Optimization Algorithm (QAOA) using hybrid Differential evolution (DE) -Trotterized Quantum Annealing (TQA) initialization method to solve the QUBO formulation of feature selection. This is a new approach to feature selection, which is very important for machine learning research. This method is applied to 11 real-life datasets and the results have been improved significantly.