论文部分内容阅读
作者用人工神经网络模型对四种细菌的DNA序列进行了分类。用“两碱基片段含量”法表征了DNA序列,将DNA序列转化为一个16维向量。然后设计了人工神经网络模型并用“留一法”进行了训练。计算结果表明:人工神经网络对所有DNA序列的分类正确率达到了84.3%,表明用人工神经网络模型可以较好地根据DNA序列的结构特征进行种类。
The authors used artificial neural network models to classify the DNA sequences of four bacteria. The DNA sequence is characterized by the “two base fragment content” method, which transforms the DNA sequence into a 16-dimensional vector. Then the artificial neural network model was designed and trained with “leaving one method ”. The results show that the correct rate of classification of all DNA sequences by artificial neural network reaches 84.3%, which shows that the artificial neural network model can be better based on the structural features of DNA sequences.