论文部分内容阅读
为了改善概率神经网络(PNN)在训练样本数量较大冗余度较高时存在的结构复杂的问题,提出一种基于主成分分析(PCA)的结构优化方法。以概率乘法公式为理论依据,根据训练样本的PCA结果对PNN进行结构优化,并引入学习算法减小PNN的参数不确定性。实验结果表明:在训练样本数量较大冗余度较高的情况下,优化后的PNN能够使用比传统PNN更简单的网络结构达到相近的结果。
In order to improve the complex structure of probabilistic neural network (PNN) when the number of training samples is high and the redundancy is high, a structural optimization method based on principal component analysis (PCA) is proposed. Based on the probability multiplication formula, the PNN structure is optimized according to the PCA results of the training samples, and the learning algorithm is introduced to reduce the parameter uncertainty of the PNN. The experimental results show that the optimized PNN can achieve similar results using a simpler network structure than the traditional PNN with a large number of training samples and high redundancy.