基于神经网络的蛋白质序列分类算法研究

摘要

To solve the difficulty of protein sequences sorting, this paper proposes a protein sequences sorting algorithm based on neural network. The characteristic parameter of protein sequences has been collected by adopting the feature extraction techniques on the basis of frequent pattern scanning. Therefore, this research builds up a neural network with three layers used for protein sequences sorting. After conducting the training to neural network by numerous data, it is learned that the algorithm designed in this paper has a high accuracy of almost 98% especially for the newly emerged protein sequences sorting.%针对蛋白质序列难以分类的问题,提出了基于神经网络的蛋白质序列分类算法,通过采用基于频繁模式扫描的蛋白质序列特征抽取技术,采集得到了蛋白质序列的特征参数.在此基础上,构建了三层的神经网络,用于蛋白质序列的分类,经过大量数据对神经网络的训练,经测试表明,文章所设计的蛋白质序列分类算法,其精度达到了98％,尤其是对于一些新出现的蛋白质序列分类效果更好.