Abstract: To enhance the clustering ability of self-origanization network, a quantum-inspired self-organization clustering algorithm was proposed based on Bloch spherical rotation. The clustering samples were mapped to the qubits on the Bloch sphere by taking all the sample values as the phases of the qubits, and the all weight values in the competitive layer were mapped to the qubits randomly distributed on the Bloch sphere. Then, the winning node was obtained by computing the spherical distance between sample and weight value, and the weight values of the winning nodes and its neighborhood were updated by rotating them to the sample on the Bloch sphere until the convergence. The obvious advantage of this method is that it has higher clustering accuracy. The clustering results of the benchmark IRIS sample show that the proposed approach is obviously superior to the classical self-organization network and the K-mean clustering algorithm.

Key words: quantum bits, Bloch spherical rotation, self-organization network, clustering algorithm