论文部分内容阅读
本文在提出规范、规范满等概念的基础上,对CC4神经网络分类计算的倾向性进行了理论分析。并针对文本分类,提出了基于神经网络的增量式索引建立方法,将以词频为基础表示的高维文本信息映射到低维数据空间。为了使CC4神经网络应用到基于文本信息空间索引的分类技术中,将空间索引变换为CCA神经网络可以接受的二值向量,使得CC4神经网络以空间索引为基础,进行文档分类。最后给出了相应的实验结果。
Based on the concepts of norm, norm specification and so on, this paper makes a theoretical analysis on the propensity of classification and calculation of CC4 neural network. And for text classification, an incremental indexing method based on neural network is proposed, which maps the high-dimensional text information based on word frequency to low-dimensional data space. In order to apply the CC4 neural network to the classification technology based on the text information spatial index, the spatial index is transformed into a binary vector acceptable to the CCA neural network, so that the CC4 neural network is classified based on the spatial index. Finally, the corresponding experimental results are given.