摘要
介绍了一种使用人工神经网络的典型模型一BP算法构造了棉花病虫害学习和诊断系统。系统性能测试结果表明该网络不仅能100%正确诊断已学样本,而且对未学样本也能以60%以上的正确率进行诊断,同时系统还有较强的容错和抗干扰能力。
In this paper, a diagnostic system based on back propagation neural network is described. The test of this system shows that it not only can fully diagnose samples which have been learned but also can diagnose samples which have not been learned in the rate of above 60%, and the same time, the system is strongly robust.The system has already been used and the users have received apparently economic results.All these results show that the diagnostic system based on neural network is better than that based on traditional expert system.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
1998年第1期71-75,共5页
Systems Engineering-Theory & Practice
关键词
病虫害
诊断
棉花
神经网络
back propagation neural network, crop disease and pests, diagnostic system.