【摘要】 与主量分析密切相关的Ｈｅｂｂｉａｎ算法大都局限于线性模型和精确输入恢复前提．本文考察了一类非线性的竞争性Ｈｅｂｂｉａｎ学习，提出了一个确切的基于纯竞争机制的优化算法．仿真实验表明，该算法较之于现有的算法更为快速、有效．更多还原

【Abstract】 Most Hebbian algorithms, closely related to principal component analysis, are subjected to restraints of linear models and the criterion of accurate input reconstruction. In this paper, a kind of nonlinear competitive Hebbian learning is examined, and an exact optimization algorithm based on pure competitive mechanism is proposed. Simulation result shows that the proposed algorithm learns faster and more effectively than previous approaches.更多还原