论文部分内容阅读
论文在SOM神经网络算法和结构进行研究的基础上,以结肠癌基因图谱信息为研究对象,根据基因特征运用T检验分析较好地剔除了数据样本中的冗余基因,从而降低神经网络的输入维数,并建立了优选基因特征对应的SOM神经网络数学模型。此方法加速网络收敛速度,从整体上提高网络的性能,从而减少样本分类错误,为临床癌症自动诊断提供了一种有效的方法。
Based on the study of SOM neural network algorithm and structure, taking the genetic information of colon cancer as the research object, using T test to analyze the gene characteristics, we can eliminate the redundant genes in data samples and reduce the input of neural network Dimension, and established a mathematical model of SOM neural network corresponding to the preferred gene characteristics. This method accelerates the speed of network convergence and improves the performance of the network as a whole, thus reducing the sample classification error and providing an effective method for the automatic diagnosis of clinical cancer.