【摘要】 文中提出了一种将GHA(Generalized Hebbian Algorithm)学习规则应用到核主成分分析的新方法,它结合了核主成分分析和GHA学习规则的优点,既能利用核主成分分析的方法方便地提取数据的非线性特征,又能避免在大样本数据的情况下运算复杂和存储空间大的问题。实验证明了该方法的可行性和高效性。更多还原

【Abstract】 Presents a new method that combines the algorithm of GHA with kernel principal component analysis which can make good use of respective advantage of two algorithms.First,this method uses the algorithm of kernel principal component analysis to extract the nonlinear feature of data.Second,it can also avoid the computational complexity and high dimensionality of space.The experiments had proved this method is feasible and efficient.更多还原