Abstract: Laboratory of Defense Technology, Shenzhen Univ., Shenzhen 518060， China)
This paper proposes a method based on quantum measurement for spherical shell clustering. This method simulates the conscious cognition of the human brain using quantum measurement. It regards a sample set as an environment quantum ensemble which is measurable and the stimulus quantum system. In this thought, each sample is treated as a microcosmic state of quantum superposition. The stimulus quantum ensemble attracts the conscious attention and makes the cognition mechanism produce corresponding spherical shell surface operators by collapsing on microcosmic states in probability. A shell surface operator is applied to measure the environment quantum ensemble and the average measure value can be obtained which validates the operator. At the same time, a sample set on the shell surface will come forth. Quantum shell clustering has the ability of perfect initialization and auto-detection of the class number of clustering compared with fuzzy shell clustering. The quantum cognition thought,which integrates unconscious data-driving and conscious concept-driving in information-processing, can be further studied and applied in the other pattern analysis.

Key words: spherical shell clustering, spherical shell surface operator, quantum measurement, quantum ensemble