神经网络识别分类中交叉模糊问题的研究

Study on Intercross Confusion in Identification and Classification of Neural Network

针对神经网络在分类问题研究中出现的交叉模糊问题 ,提出误差比对法对模糊问题加以解决。以BP网络对TBM和诱饵进行分类识别为例 ,分析表明该方法精度可靠 ,可望较好地解决分类问题中的模糊问题。

In this paper,a new theory called the theory of error contrast is put forward to solve the question of intercross confusion in identification and classification of neural network The result of analysis shows this method is accurate and reliable through the example of identification of TBM and bait based on BP network This method can solve the fuzzy question in classification of neural network