摘要
针对传统大坝安全监测预报模型存在的受环境量相关性影响较大和网络训练时间长、易收敛到局部最优点等缺点,通过对原始监测信息的粗糙集预处理提取主要影响因素和决策规则集,并通过对规则集的不确定性推理建立了大坝监测的粗集预报模型.与传统方法相比,粗集预报模型不依赖于具体的数学模型,所得到的预报值是一定可信度下的区间值,并且预测值区间与实测值区间基本一致.
In consideration of the fact that the conventional prediction model for dam safety monitoring is easily affected by the correlation of environmental factors, and the fact that it needs a long time of network training and easily converges to local optimum points, a rough set prediction model is developed by extraction of main affecting factors and decision rules from pretreatment of rough sets of original monitoring data and reasoning of decision rules under uncertainty. Compared with the conventional method, the present model is independent on a special mathematical model, and the predicted results are intervals under certain reliability, which are in accordance with measured data.
出处
《河海大学学报（自然科学版）》
CAS
CSCD
北大核心
2005年第4期391-394,共4页
Journal of Hohai University(Natural Sciences)
基金
国家自然科学基金重点资助项目(50139030)
国家"973"计划资助项目(2002CB412707)
关键词
大坝安全监测
粗集理论
不确定性推理
区间预报
<Keyword>dam safety monitoring
rough set theory
reasoning under uncertainty
interval prediction
