关键词: 独立分量分析, 人工蜂群, 峭度, 精度

Abstract: For the fast fixed point independent component analysis method is easy to fall into local optimal solution, an improved ICA method based on artificial bee colony algorithm is proposed. The algorithm uses the absolute value of kurtosis as the optimized objective function, improves the following and scouts stage searches equation of artificial bee colony algorithm and optimizates the independent component analysis. At the onlookers stage, the new candidate positions are generated near the optimal solution in the current iteration, to avoid weakening the exploitation ability with random search. At the scouts phase, the distance between the optimal and the worst solutions of the current iteration is utilized to replace the worst solution in the process of optimization, so as to improve the optimization effect of artificial bee colony optimization method and the accuracy of independent component analysis. The experiment confirms the effectiveness of the improved algorithm.

Key words: Independent Component Analysis（ICA）, artificial bee colony, kurtosis, accuracy