【摘要】 介绍了 Probabilistic Neural Networks(PNN)网的结构和算法 ,给出了点积和生长等相关概念的定义 ,研究了它们的性质 ,并用生长的方法改造训练样本集以便提高识别精度 .在模拟实验中 ,直接使用 PNN的算法 ,不能对水泥强度测试样本正确分类 ,使用本文提出的方法 ,能够对水泥强度的测试样本正确分类 ,识别率达到 10 0 % ,从而显示本文提出的方法是可行的和非常有效的更多还原

【Abstract】 Described a topology and an algorithm of Probabilistic Neural Networks. The dot product, growing vector, growing factor and other terms had been defined, their characters were studied. A method to improve the accuracy of Probabilistic Neural Networks was proposed. In the simulation experiment, the algorithm of PNN can not recognize the cement strength testing samples perfectly, the proposed algorithm can identify the cement strength testing samples perfectly, the recognition rate is 100%. The simulating result showed that the method is feasible and efficient.更多还原