论文部分内容阅读
本文对用线性神经网络提取主分量的现有各种算法和作者提出的算法进行了分析和实验比较。模拟实验结果表明,所提出的算法以收敛的单调性、收敛速度较快、收敛精度较高而优于其它算法。
In this paper, we analyze and compare the existing algorithms for extracting principal components using linear neural networks and the algorithms proposed by the authors. Simulation results show that the proposed algorithm is better than other algorithms in terms of convergence monotonicity, faster convergence rate and higher convergence accuracy.