基于RBF神经网络的水泥强度预测模型毕业设计.doc

基于RBF神经网络的水泥强度预测模型 摘 要 水泥强度是指水泥试件单位面积上所能承受的外力，它是水泥的最重要性能。28天龄期是强度基本稳定的龄期，国内外都把水泥的28天强度作为通用水泥的代表强度，所以测定水泥强度值的主要依据是水泥的28天抗压强度。神经网络具有很强的学习、联想和容错能力及高度非线性函数映射能力，粗糙集可以从数据中挖掘有用的知识，因此本文采用粗糙集和RBF神经网络建立水泥强度的预测模型。首先利用模糊c均值聚类算法对连续数据集进行离散化。然后利用粗糙集对数据集进行属性约简，提取出影响水泥28天抗压强度的主要因素。根据以提取的主要因素作为RBF神经网络的输入变量建立预测模型。 通过仿真实验对建立的预测模型进行验证，实验结果表明本文建立的RBF神经网络模型具有较高的预测精度。神经网络为水泥强度的快速预测提供了一种新方法，有较好的实用价值。 关键词: 粗糙集 RBF神经网络 水泥28天抗压强度 预测模型 Cement strength Forecast Model Based on RBF Neural Network Abstract The strength of cement is the cement specimens per unit area can withstand external forces, it is the most important properties of the cement. 28 days of age as the strength of the basic stability of age. 28-day strength of cement and concrete at domestic and abroad as a general representative of cement strength, The neural network has a strong learning, Lenovo and fault-tolerant and highly nonlinear function mapping capability, Rough sets in data mining of useful knowledge This article uses the rough set and RBF neural network to establish a cement strength prediction model First the use of fuzzy c-means clustering algorithm for discretizational of continuous data sets. Then use the rough set attribute data set reduction, extraction of 28 days compressive strength of cement. According to extraction of main factors of establishment as input variables of the RBF Neural network forecast model. Verified through simulation experiments to establish the neural network model, the experimental results show that the article uses RBF neural network can greatly improve the prediction accuracy of the strength of cement. The neural network provides a new method for the rapid prediction of cement strength, has the good practical value. Key Words: Rough Sets; RBF Neural network; Cement compressive strength of 28 days; Forecast model. 目 录 基于RBF神经网络的水泥强度预测模型 I 摘 要 I Abstract II 目 录 III 第一章 绪论 1 1.1课题研究的背景和意义 1 1.1.1 水泥的工艺流程 1 1.1.2 课题研究的意义 3 1.2 课题的任务及解决方法 3 1.3 论文的结构 4 第二章 课题相关的理论基础 5 2.1连续属性离散化 5 2.1.1离散