从受训神经-模糊模型中提取知识

本文对用于模式分类、函数逼近、参数估计的多层感知器 (MLPs)给出 1个清晰的关于内部行为的解释。作者以单隐层的 MLP为例 ,论述了关于 MLP的内部行为的半线性分析理论。对受训的MLP,将隐层单元的输出分别定义为网络输出的正、负“内部分量”;定义内部分量的连接权重集为给定问题的“内部判别模式”;建立了 MLP和模糊集相结合的新模型 ;分析了 MLP的结构为 N- 2 - 1和N- H- 1 ,给出权重初始化的方法 ;提出了 1种从受训神经 -模糊模型 (NFMs)中提取知识的全新的具有实用价值的方法。     

Shoreline variability via empirical orthogonal function analysis: Part I temporal and spatial characteristics

Empirical orthogonal functions (EOFs) or principal components were used to extract the significant modes of shoreline variability from several data sets collected at three very different locations. Although EOFs have proven to be a valuable tool in the analysis of nearshore data, most applications have focused on the ability of the technique to describe cross-shore or profile variability. Here however, EOFs were used to help identify the dominant modes of longshore shoreline variability at Duck, North Carolina, the Gold Coast, Australia, and at several locations within the Columbia River Littoral Cell in the U.S. Pacific Northwest. In part one of this analysis, characteristic patterns of shoreline variability identified by the EOF analysis are described in detail. At each site, the dominant modes consisting of the first four eigenfunctions were found to describe nearly 95% of the total shoreline variability. At both Duck and the Gold Coast, several interesting longshore periodic features suggestive of sand waves were identified, while boundary effects related to natural headlands and navigational structures/entrances dominated the Pacific Northwest data sets.     

ASM1中污水水质参数与化学计量系数的估测

简要介绍了活性污泥1号模型中14个污水水质参数与5个化学计量系数的估测方法和典型取值。     

绿片岩三轴流变力学特性的研究(II):模型分析

首先,基于在岩石全自动流变伺服仪上得到的绿片岩三轴流变试验曲线,采用五元件线性粘弹性模型对表现为粘弹性流变特性的曲线进行了辨识,获得了绿片岩的粘弹性流变参数;然后,提出了一个新的非线性粘性元件,并将其与塑性体并联起来,得到一个新的非线性粘塑性体(NVPB),该体能充分反映岩石的加速流变特性:同时,将NVPB模型与五元件粘弹性模型串联起来,建立了一个新的岩石七元件非线性粘弹塑性流变模型。采用绿片岩加速流变全过程曲线,对提出的岩石七元件非线性粘弹塑性流变模型进行了辨识,得到了岩石七元件非线性粘弹塑性流变模型的材料参数。流变模型与试验结果的比较,显示了所建模型的正确性与合理性。     

长治市防雹作业新体系

本文根据长治市多年冰雹发生情况，经过分析统计，找出了冰雹多发源地、雹云发晨方向、路径和影响范围，提出了合理布局，联防作业，建立现代化作业指挥系统的防雹新思路。     

城市化对北京平均气温的影响

本文利用北京（观象台）及周围几个气象台站月平均气温资料，滑动平均后，采用主成分分析方法，再进行倾向性分析，探讨了北京城市化对气平均气温的影响，估算对北京（观象台）气温记录的影响约为０．２１℃／３３ａ推断市中心二环路以内强烈地影响。     

DYNAMIC VARIATIONS OF WATER QUALITY IN TAIHU LAKE AND MULTIVARIATE ANALYSIS OF ITS INFLUENTIAL FACTORS

ＬａｋｅＴａｉｈｕｉｓａｌａｒｇｅｓｈａｌｌｏｗｌａｋｅｉｎＣｈｉｎａ,ｗｈｉｃｈｐａｌｙｓａｎｉｍｐｏｒｔａｎｔｒｏｌｅｉｎｗａｔｅｒ-ｑｕａｎｔｉｔｙｒｅｇｕｌａｔｉｏｎ,ｗａｔｅｒｓｕｐｐｌｙ,ｃｕｌｔｉｖａｔｉｏｎｅｔｃ.Ｗｉｔｈｒａｐｉｄｄｅｖｅｌｏｐｍｅｎｔｏｆａｇｒｉｃｕｌｔｕｒｅａｎｄｉｎｄｕｓｔｒｙｉｎｔｈｅａｒｅａ,ａｌａｒｇｅｎｕｍｂｅｒｏｆｐｏｌｌｕｔａｎｔｓｈａｖｅｂｅｅｎｄｉｓｃｈａｒｇｅｄｉｎｔｏｔｈｅｌａｋｅ,ｌｅａｄｉｎｇｔｏａｎｉｎｃｒｅａｓｅｉｎｎｕｔｒｉｅｎｔｓ…     

Integrated spatial sampling modeling of geospatial data

Sampling is to, by efficient selection of samples, acquire the accurate information about the population (the research object) at less cost. Spatial sampling is a kind of sampling toward geospatial objects or features with spatial correlation. The differences between effi-cient sampling and completely universal survey lie in quality, time and cost. Sampling provides a kind of economical, prompt and accurate survey[13]. Efficient spatial sampling can be regarded as the optimization of the sampl…     

鲁西南深埋粘性土物理力学指标多元统计研究

运用R因子分析和R型聚类两种多元统计分析方法,对鲁西南地区深埋粘性土的20组样的8个物理力学指标数据进行了分析。利用R因子分析方法,揭示了各指标对鲁西南地区深埋粘性土状态和力学性质的影响关系。利用R型聚类分析方法,研究了深部粘性土各物理力学指标之间的相关性,为进一步从物理力学指标对深埋粘性土的特殊状态和力学特性的研究打下基础。     

基于方差分量估计的自适应融合导航

多源传感器观测信息的精度一般不一致,利用方差分量估计理论可以合理地顾及各传感器观测信息在融合导航解中的权比.首先给出基于多源观测信息的融合导航解算模式,进而基于方差分量估计讨论了各传感器观测信息在融合导航解中的合理平衡问题,然后应用自适应因子顾及动力学模型预报信息的贡献,给出了计算过程,并利用模拟数据进行了试算与比较.