DBL(1,0,1)误差非线性回归模型相关性和方差齐性的检验

讨论了具有双线性DBL(1，0，1)误差的非线性回归模型的相关性和方差齐性的检验问题，用Score检验方法给出了双线性项检验以及相关性和方差齐性同时检验的检验统计量，并用随机模拟验证了检验方法的正确性；推广和发展了具有线性序列误差回归模型的结果；最后将结果应用于DBL(0，1，1)误差的非线性回归模型。     

基于支持向量机的短期GDP预测模型与应用研究

支持向量机是最近几年国际上模式识别研究的热点,具有全局最优和良好的泛化能力.本文在理论分析江门市GDP预测指标体系的前提下,研究了基于SVM的预测方法,并运用实际数据进行建模和预测,获得了比较准确的预测结果.     

向量自回归模型参数估计矩阵的渐近分布

讨论了向量自回归模型参数的估计矩阵П^、Ω^的渐近分布，给出并证明了两个相应的结论．     

基于灰色理论的煤炭需求预测模型研究

研究了GM(1，1)模型在我国煤炭需求预测中的应用，并以实际数据为基础，建立了我国煤炭需求量的数列预测模型。经检验，模型可靠，可用于对我国煤炭需求总量的预测。简要分析了根据实际变化不断改进模型的必要性。     

基于最小方差控制的闭环辨识信号设计

讨论了系统辨识实验信号的设计方法．从实际控制器和理想控制器间误差的角度出发，评价系统性能指标为系统在实际和在理想控制器控制下输出误差的平方均值最小．当输入信号或输出信号能量有限时，分别推导出辨识输入信号功率谱密度．当控制策略为最小方差控制时，给出信号功率谱的具体表达式．结论对在最小方差控制下的闭环辨识信号设计有指导作用．     

不同类型水稻两用核不育系配组F1代产量性状的比较研究

通过利用4个常用两用核不育系与10个恢复系杂交．对F1代产量性状进行差异性分析、一般配合力比较、相关性分析和通径系数比较．结果表明．不同类型不育系配组后代性状中．均有不同的优势表现．对产量的主导影响因子不同．一般配合力表现差异较大．且对产量的相对影响重要性不同．在配制组合中，要综合兼顾各性状间的关系．不同不育系不同对待。     

人工神经网络构造经济预测模型方法的探讨

探讨利用人工神经网络建立经济模型的一般原理和方法,并通过构造一个宏观经济模型的实例加以说明.与传统的回归分析方法相比,人工神经网络具有良好的适应性和分析预测能力.     

基于BP神经网络预报的动态矩阵预测控制

提出了一种基于BP神经网络预报的动态矩阵预测控制新算法 ,在该算法中 ,先用BP神经网络辨识对象模型 ,同时预测对象的未来输出 ,然后用动态矩阵控制算法进行滚动优化和反馈校正。该方法解决了非线性、时变对象难以建模及控制的问题 ,仿真结果验证了这一新型算法的可行性     

A multi-layer zone model for predicting temperature distribution in a fire room

Modelingoffireinacompartmentcanbeachievedeitherusingazonemodelingmethodoracomputationalfluiddynamics (CFD )modelingmethod .Themostcommonzonemodelisthetwo zonemodel.Themaincharacterofthismodelisthatitdividestheroom (s)intoahot ,upperlayerandacooler,lowerlayer ,andthatthephysicalpropertiesofeachlayer ,suchasitsgastemperatureandspeciesconcentrationsarealluniform .Somemodelsofthistypehavebeendeveloped[1～3] ,andacomprehensivereviewoftheexistingtwo zonefiremodelscanbefoundinreference [4 ].Thecomp…     

Approaches to prediction of impact of Qinghai-Tibet Railway construction on alpine ecosystems alongside and its recovery

With the aid of the Remote Sensing (RS) and Geographic Information System (GIS) technology, the ecosystem pattern and fragility distribution maps of the 50-km-wide zone along the Qinghai-Tibet Railway were compiled and by using the superimposition method, range, area and indexes of the impact of various engineering activities on the ecosystems alongside the railway were studied. By making reference to the ecosystem recovery process of the Qinghai-Tibet Highway, mechanisms of recovery of the alpine ecosystems alongside the Qinghai-Tibet Railway were studied and extents and rates of the recovery were predicted.The results indicate that the impact of the railway engineering on the Alpine ecosystem depends mainly on how much the original surface soil in the zone has been disturbed and how fragile of the ecosystem per se. Restoration of vegetation coverage and species abundance shows a significantly reverse relationship with disturbance of the original surface soil but an extremely positive one with the length of the restoration period and mean annual precipitation and annual mean relative humidity in the period and no obvious bearings with altitude and temperature. In sections with an annual precipitation over 200mm, as long as a certain percentage of original soil is left in situ, it takes only 30 years or so for biodiversity to get basically restored to the original level after the construction is completed but at least 45-60 years or more for vegetation coverage.