Abstract: To improve computation speed and efficiency of Genetic Algorithm-Partial Square Least (GA-PLS), a novel feature selection algorithm which combines quantum computation and GA-PLS (QGA-PLS) is proposed. In QGA-PLS algorithm, qubits and superposition of states are used for chromosome code. Quantum rotation gate is used for genetic operation to update parameters and enhance population diversity. Meanwhile, with PLS model which is reconstructed by quantum computing, the value of individual adaptability is calculated. Rapid convergence and good global optimization capability characterize the performance of QGA-PLS. The proposed method is applied to two simulation experiments, extreme value of a function and feature selection for Iris dataset. The experimental results indicated that, compared with QGA and GA-PLS, QGA-PLS has better performance in feature selection, execution time and classification accuracy, which proves the efficient of proposed method.

Key words: quantum optics, quantum genetic algorithm-partial square least (QGA-PLS), quantum computation, feature selection