排序方式： 共有8条查询结果，搜索用时 15 毫秒
1
1.
小波网络方法用于时间序列分析中非线性模型的研究 总被引：1，自引：0，他引：1
把小波网络方法用于时间序列分析中非线性模型，利用小波网络的非线性逼近能力，构造一个非线性函数，并对温度资料进行处理，得到了较好的拟合效果。 相似文献
2.
本文把1958—1984年天山乌鲁木齐河源1号冰川的融水径流总量的时间序列X(t),分解为趋势项L(t)、周期项P(t)、平稳项S(t)、随机项ε(t),使该时间序列表示为X(t)=L(t)+P(t)+S(t)+ε(t)。采用非线性回归提取L(t),用谱分析和Fourier级数提取P(t),余差用自回归方程建模,用上述项的叠加作出预报,按相关指数公式计算R=0.90,效果良好。 相似文献
3.
为了使大坝变形的预测精度更高,针对大坝形变量的时间序列中存在着非平稳和非线性等曲线特性,使用一种经验模态分解(EMD)和非线性自回归动态神经网络(NAR)相结合的EMD-NAR模型对大坝形变时间序列进行预测。以某大坝实测的时间序列数据为算例,分别使用BP模型、NAR模型和EMD-NAR模型进行实验对比,结果表明,BP、NAR、EMD-NAR模型预测的均方根误差(RMSE)分别为0.9449,0.6993,0.4678;模型预测的平均相对误差(MRE)分别为0.1492,0.1065和0.0688,从三种模型预测结果对比可知,组合的EMD-NAR模型预测精度最高且稳定性最好,为时间序列的大坝形变预测提供一种新的参考思路。 相似文献
4.
本文把1958—1984年天山乌鲁木齐河源1号冰川的融水径流总量的时间序列X(t),分解为趋势项L(t)、周期项P(t)、平稳项S(t)、随机项ε(t),使该时间序列表示为X(t)=L(t)+P(t)+S(t)+ε(t)。采用非线性回归提取L(t),用谱分析和Fourier级数提取P(t)余差用自回归方程建模,用上述项的叠加作出预报,按相关指数公式计算R=0.90,效果良好。 相似文献
5.
Diel fluctuations in discharge not associated with precipitation have been identified in streams around the world and attributed to evapotranspiration of both hillslope and near-stream vegetation. Several mechanisms have been hypothesized to explain the generation of these signals, but a consensus has not been reached. In order to investigate the origin of these daily discharge fluctuations, we measured discharge and groundwater levels in a 43-ha headwater catchment near Cullowhee, NC, USA. The catchment was logged at least twice in the past 120 years, allowing the evergreen invasive Chinese privet (Ligustrum sinense) to become established in near-stream areas. Stream discharge was measured at the outlet, and stream stage and near-stream groundwater level were measured at a second location higher in the catchment. Daily fluctuations in discharge were quantified and found to increase in magnitude with daily mean discharge. Interestingly, daily discharge fluctuations were observed in the dormant season, and our analysis suggests that their magnitude may increase more with discharge than during the growing season. Two previously hypothesized processes may explain our observations. A decrease in signal dampening due to faster transmission of an evapotranspiration signal at elevated flows may explain the increased diel fluctuation magnitude at higher flows. Additionally, because the primary active vegetation in our catchment in the dormant season is in near-stream areas, we suggest that this indicates most of the evapotranspiration signal in stream discharge is derived from near-stream vegetation, not hillslopes. 相似文献
6.
针对天山中段地区水平地壳形变特征最新动态的研究问题,提出了利用GPS连续站及流动站资料,基于速度场结果进行GPS剖面分析的方法来进行研究。结合速度场及剖面结果可以看出,北西走向的喀什河断裂带,1992—2011背景走滑运动为右旋走滑,2009—2014年右旋运动有所减弱,2013—2015年仍然较弱;天山中部的滑脱断层,1992—2011年的结果显示其走滑运动较不明显,但2009—2014期的结果表明出现了一定的左旋运动,2013—2015年的左旋运动加剧;天山中部的滑动速率结果显示出天山中段博罗科努断裂带中段变形速率较东西两侧均较小,值得进一步关注。研究结果利用了最新数据,显示了天山中段水平地壳形变的最新特征,为判定天山中段地区的地震危险性提供了支持。 相似文献
7.
The dynamics of the dunefoot along a 160 km portion of the Dutch coast has been investigated based on a data set of annual surveys dating back to as early as 1850. The linearly detrended (or residual) dunefoot positions comprise an alongshore uniform and an alongshore non‐uniform component. The former is expressed as 10 to 15 m of landward retreat along extensive (>10 km) stretches of coast during years with severe storm surges and as up to 5 m of seaward advance during years without significant storm activity. The latter, alongshore non‐uniform component is organized in sandwave‐like patterns, which may have a longevity of decades to up to the duration of the entire data set (150 years). Their wavelengths vary along the coast, from 3·5 to 10 km; migration rates are 0–200 m a?1. Dunefoot sandwaves are shown to be the shoreward extensions of similar sandwave patterns in the beach position. The non‐uniform dunefoot behaviour constitutes at least 80 per cent of the total residual dunefoot dynamics, implying that along the Dutch coast residual dunefoot variability is controlled by temporal and spatial variability in beach characteristics, and not by storm‐induced uniform erosion. Various potential mechanisms causing beach sandwaves are discussed. Copyright © 2002 John Wiley & Sons, Ltd. 相似文献
8.
针对现有出租车轨迹数据挖掘中时间序列邻近度量方法存在的问题,提出一种基于DBSCAN算法和改进的DTW距离的时间序列聚类算法提取具有相似性出行特征的时空模式,进而研究城市人群出行行为的时空差异。以南京市为例,结合电子地图对出行模式的空间分布特征进行分析,证明了本文所提出的方法的有效性。实验结果表明:在空间分布上,工作日出租车出行模式按照平均出行频次由高到低排序,从城市中心向四周扩散,呈中心环状分布,出行模式区域界限较为明显,同类出行模式分布区域对应相似的功能。提出了一种基于DBSCAN算法和改进的DTW距离的时间序列聚类算法提取具有相似性出行特征的时空模式,有效地分析城市人群出行行为的时空差异。 相似文献
1