论文部分内容阅读
提出将改进的BP神经网络应用于森林防火专家系统的不确定性推理中,其良好的自学习和泛化能力,可以解决基于可信度规则的知识表示在实际应用中导致的规则激增,推理速度缓慢的问题。该方法将不确定的知识用可信度区间表示,通过知识编码,设计并训练BP网络,最后用MATLAB进行仿真。实验结果表明:BP神经网络可以自动学习专家的典型经验,并且能将之准确的推广,隐含层神经元个数的确定和典型样本的选取决定了准确精度。在实际的专家系统不确定推理应用中具有应用价值。
Proposed to apply the improved BP neural network to the uncertainty of forest fire expert system, its good ability of self-learning and generalization can solve the rule-based explosion caused by the knowledge representation based on the credibility rules in practice, reasoning Slow speed In this method, the uncertain knowledge is represented by the confidence interval, and the BP network is designed and trained by knowledge coding. Finally, the simulation is carried out by using MATLAB. The experimental results show that BP neural network can automatically learn the typical experience of experts and can be accurately generalized. The number of hidden neurons and the selection of typical samples determine the accuracy. It has application value in practical application of uncertain system of reasoning.