利用模糊神经网络直接从实验数据中提取规则 ,进行材料性能建模与预测 ,作为应用示例 ,建立了基本成分和组织参数的灰铁预测模型。与多元统计分析、模糊回归和广义回归网络所得的结果相比 ,该方法所得的模型具有学习精度高 ,且具有更好的泛化能力。 The fuzzy neural network is used to directly extract rules from the experimental data to model and forecast the material properties. As an application example, the gray iron prediction model with the basic composition and the organizational parameters is established. Compared with the results obtained from multivariate statistical analysis, fuzzy regression and generalized regression network, the model obtained by this method has high learning accuracy and better generalization ability.