摘要
BP神经网络具有并行处理、自适应学习、容错能力好等优点,根据地下燃气管道安全评估的应用要求,对其算法进行改进,结合专家系统的设计思路,构造了一个基于BP神经网络的地下燃气管道安全评估专家系统,包括知识表示、知识获取和推理解释等模块。然后利用已有的样本编制软件进行实验,结果表明,该系统评估正确率达91%,推理效率高、扩展性好,从而很好地满足了实际应用的要求。神经网络为专家系统的发展提供了一条新的途径。
BP neural network has many advantages such as parallel scheduling, self-adapt learning, anti-interference capability. According to the characteristics of the security evaluation expert system of underground gas conduits, BP algorithm is improved, then an expert system based on BP neural network is built, which includes knowledge representation, knowledge acquisition, inference engine and so on. Then the system is realized using the samples. The expert system, whose accuracy is 91%, has high effective, inference engine, better expansibility. Neural network provide a new approach to improve expert system.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第2期425-426,439,共3页
Computer Engineering and Design
基金
教育部人文社会科学研究规划基金项目(2005-241)
广东省科技攻关基金项目(2005B10101010)
关键词
BP
神经网络
专家系统
燃气管道
安全评估
BP
neural network
expert system
gas conduits
security evaluation
