基于知识神经网络的大肠杆菌启动子识别算法的改进及C++实现  

Improvement of Algorithm Based on the KBANN for Recognizing E.coli Promoters and Realized by C++ Program

作　　者：周晖杰[1] 史定华[1] 

机构地区：[1]上海大学理学院数学系,上海200436

出　　处：《应用数学与计算数学学报》2004年第2期1-7,共7页Communication on Applied Mathematics and Computation

基　　金：国家863资助项目功能基因组的信息分析(2002AA234021)重点项目的支持.

摘　　要：本文简要介绍了基于知识神经网络算法在启动子识别中的应用以及对该算法的若干改进,并通过C++实现.对86个启动子与82个非启动子样本,利用改进算法并通过交叉验证,其预测结果显示误分类样本个数仅为2个,达到了更高的准确率.In this paper, we briefly describe KBANN (Knowledge Based Artificial Neural Network advanced by Shavlik & Towell, 1989) for recognizing E. coli promoters. Then we have been made several improvements of this algorithm and realize it by C++ program. We select 86 promoters and 82 non-promoters as samples, and the results show that the number of misclassified samples is two and it is better than before (In the Shavlik and Towell's paper, the number of misclassified samples is four). Apparently, the results show the feasibility and validity of the improved algorithm by cross validation method.

关 键 词：C++ 基于知识 识别算法 神经网络 改进算法 显示 交叉验证 识神 启动子 大肠杆菌 

分 类 号：O235[理学—运筹学与控制论]