关键词: 特征值分解, QR分割, 谱聚类算法, 量子遗传谱聚算法

Abstract: On clustering large data, the spectral clustering can hardly be promoted because of large occupied storage space and high time complexity. Hence, multi-segment and upward and downward double-direction shrink QR algorithm for the corresponding eigenvectors of eigenvalues was adopted to achieve dimensionality reduction. Then, a new quantum genetic spectral clustering algorithm was proposed to cluster the sample points in the mapping space. Compact input with low-dimension for quantum genetic spectral clustering was obtained after mapping, and the quantum genetic spectral clustering algorithm, characterized by its rapid convergence to global optimum and minimal sensitivity to initialization, can obtain good clustering results. The experimental results show that the proposed method is superior to the spectral clustering algorithm, K-means, NJW algorithm in astringency and stability and has a higher overall optimal solution.

Key words: eigenvalue decomposition, QR segmentation, spectral clustering algorithm, quantum genetic spectral clustering algorithm