摘要
大多分类问题由于各特征之间存在交互性等不确定性问题,使一般的神经网络分类算法难以取得好的分类效果。该文通过构造新特征,优化了神经网络的输入,并给出一个优化的基于神经网络的分类算法。文章还引入了分类问题的离散度的概念。并结合实例进行分析。
A general classification algorithm of neural networks is unable to gain satisfied results because of the uncertain problems existing among the features in most classification programs ,such as interaction.With new features constructed,the input of neural networks is optimized and an optimized classification algorithm based on neural networks is offered.A concept of dispersion of a classification program is introduced too in this paper.At the end,an analysis is made with an example.
出处
《计算机工程与应用》
CSCD
北大核心
2002年第5期71-73,共3页
Computer Engineering and Applications