模型输入对模糊神经网络预报模型的影响研究

为了探索模型输入对模糊神经网络预报模型预测性能的影响,研究了通过减少预报模型自变量组合的复共线性影响,并结合相似系数计算分析方法建立了一种新的模糊神经网络预报模型。以气象学科的逐日降水预报作为研究对象,利用这种新的模糊神经网络预报模型进行了实际预报试验,并与常规的模糊神经网络预报方法,中国气象局T213数值预报模式以及逐步回归预报方法的预报结果进行了对比分析。结果表明,这种基于条件数和相似系数计算的模糊神经网络预报新方法对49天降水的独立样本预报平均绝对误差为7.33 mm,预报误差比模糊神经网络预报模型下降了5.9%,比传统的逐步回归方法下降了14.9%,比中国气象局T213数值预报模式的预报结果下降了13.4%。显示了很好的应用前景。     

基于相似离度的模糊神经网络方法在降水预报中的应用

针对样本矩阵中的奇异样本可能会影响模型的预报能力，提出了一种基于相似离度的模糊神经网络模型，并对广西北部前汛期（5月～6月）逐日平均降水量进行了49天的业务预报试验。结果表明，基于相似离度的模糊神经网络模型计算速度快，预报结果与实际值符合的较好，具有较好的业务应用前景。     

基于粗糙集的模糊神经网络降水预报模型研究

为了提高模糊神经网络预测模型的预测能力,提出了基于粗集理论方法选择自变量作为模型输入的模糊神经网络预报建模方法.以短期降水预报作为研究对象,利用粗集理论的条件属性约简计算分析方法,对初选得到的预报因子矩阵进行属性约简,剔除不相关的属性,找出与预报量直接相关的预报因子,建立模糊神经网络的降水预报模型.实际的预报试验结果表明,该预报方法的预报精度明显高于由逐步回归方法选择预报因子作为模型输入的模糊神经网络预报模型及中国气象局T213数值预报模式的预报结果.     

模糊神经网络区域降水预报系统

采用模糊神经网络方法结合粗糙集理论和主成分分析技术,研制了模糊神经网络区域降水预报系统;系统利用VB和MATLAB6.5联合开发,并采用中文菜单显示的人机对话方式操作,使模糊神经网络的网络结构和多种参数的调整变得灵活而易干操作;实际应用表明,该系统对广西3个区域平均降水预报效果良好,适用性强,为广西有效防御和减少灾害提供了新的有效方法和业务服务系统.     

基于动态K均值的RBF神经网络日降水预报模型

针对优化径向基函数神经网络的各参数问题,提出一种动态K均值混合优化RBF神经网络并应用于广西降水数据进行建立预报模型,该模型与传统的K均值RBF模型和同期的T213降水预报进行对比,结果表明。该模型建立的5月3个区域的逐日降水预报,预测的精确度明显高于同期的T213降水预报。     

实现准确预报的降水预报模型构建方法研究

研究降水预报模型的构建问题,提高预报准确度.在降水预报技术中,选用较多的预报因子构建预报模型并根据预报因子之间的联系训练得到准确模型,但受空气不稳定特性的影响使得预报因子之间的非线性联系极难准确描述,传统预报模型构建方法不能有效获取预报因子之间的联系,无法训练得到准确的预报模型,造成降水预报准确度不高的问题.为解决上述问题,提出模糊聚类算法构建降水预测模型的方法.根据空气的流动特性,采用模糊聚类算法分析预报因子内部的直接关联特性从而准确表述预报因子之间的联系,构建初始预报模型,并据最小二乘回归方法训练得到降水预报模型.实验表明,模糊聚类算法能够有效获取预报因子之间的联系,准确构建和训练预报模型,实现了降水的准确预报.     

人工神经网络在小区域强降水预报应用研究

针对小区域强降水的非线性性质,利用T213数值预报产品,通过人工神经网络建模方法进行预报释用,对数量众多预报因子采用经验正交分解方法,浓缩大量因子的有效信息,建立逐日小区域强降水的人工神经网络预报模型.运用与实际业务预报相同的方法进行逐日预报试验,并与回归预报模型进行比较.结果表明,人工神经网络预报模型对小区域强降水的TS评分为0.67,而逐步回归模型的TS评分仅为0.20.由此可见,人工神经网络具有较强的处理非线性问题能力,在小区域强降水应用中具有更好的预报效果.     

模糊神经网络用于SARS疫情系统的辨识和预报

提出了模糊神经网络用于SARS疾病疫情非线性系统建模和预报的思想,该方法可以推广到各种流行性疾病的预防和控制中.模糊神经网络主要应用于非线性系统的建模、预报和控制,特别适合于不同输入类型的模型系统.而流行性疾病的传播规律与模糊神经网络模型特点相符合,这里提出将模糊神经网络用于SARS疾病疫情非线性系统的辩识和预报的观点,相应的也可推演到其它流行性疾病传播规律中.     

无线局域网业务流预报方法比较

介绍了基于时间序列、神经网络和小波的多种网络业务的预报方法,应用真实的无线局域网业务流序列检验了这些模型的预报性能,结果表明,和其他预报模型相比,基于神经网络的模型能够比较精确地捕获无线局域网业务流自身的特性,对业务流具有良好的预报性能,而基于ARIMA模型的预报性能最差。     

模糊神经网络用于SARS疫情系统的辨识和预报

提出了模糊神经网络用于SARS疾病疫情非线性系统建模和预报的思想,该方法可以推广到各种流行性疾病的预防和控制中。模糊神经网络主要应用于非线性系统的建模、预报和控制,特别适合于不同输入类型的模型系统。而流行性疾病的传播规律与模糊神经网络模型特点相符合,这里提出将模糊神经网络用于SARS疾病疫情非线性系统的辩识和预报的观点,相应的也可推演到其它流行性疾病传播规律中。     

聚类在风廓线雷达数据特征分析中的应用

认识风廓线雷达数据在不同天气条件下的普适性特征对于改善预报服务质量具有重要意义.本文基于数据挖掘中的聚类技术,以天津静海风廓线雷达逐时观测数据为研究对象,构建了风廓线雷达数据特征聚类分析模型,并在此基础上挖掘出了静海风廓线雷达的最大探测高度、最大垂直速度在晴天、多云、降水发生前、降水期间和降水结束后等天气条件下的不同特征,为天气预报服务提供了新参考依据,为风廓线雷达数据特征分析提供了新思路.     

沈阳市降水概率预报模式研究

利用日本数值天气预报产品和地面天气图，集成多种统计学方法对沈阳市降水概率进行预报，经专家订正后对社会发布。本预报模式避免了单一统计学方法的局限性，并且实现了历史因子的自动追加，能够反映最新的天气信息。经在沈阳市气象台一年的试预报，根据预报结果计算得半贝里尔得分B=0．065；Bias得分为Bias=2％；预报技巧得分Bs=80．8％。同时准确率明显高于日本数值预报结果。     

集对分析在系统智能预测中的应用综述

凡事预则立,不预则废.但事物的预测面临不确定性干扰.本文综述集对分析理论在天气降水预报、沙尘暴预报、水文水资源和供需水预测、电力与能源预测、地质灾害预测、民航风险与事故预测、作物产量预测、流脑预测、社会经济预测等方面的应用,并把基于集对分析理论的系统智能预测建模基本步骤归纳为3步.首先,构造集对并分析集对中两个集合的全...     

基于网格的中尺度数值天气预报系统设计与实现

针对国家高性能计算环境网格计算的特点，在长沙网格点上实现了一个包括全球中期数值预报、有限区域数值预报，数值预报产品释用和五维数据可视化相配套的高分辨率中尺度数值天气预报系统。该系统的预报区域可以移动，网格可以加密，适合于区域和省一级气象中心作中尺度数值预报业务试验和尺度数值预报科学研究。介绍了该系统的组成、区域同化方案以及网格计算实现技术。     

Application of fuzzy time series models for forecasting pollution concentrations

In the paper a model to predict the concentrations of particulate matter PM10, PM2.5, SO₂, NO, CO and O₃ for a chosen number of hours forward is proposed. The method requires historical data for a large number of points in time, particularly weather forecast data, actual weather data and pollution data. The idea is that by matching forecast data with similar forecast data in the historical data set it is possible then to obtain actual weather data and through this pollution data. To aggregate time points with similar forecast data determined by a distance function, fuzzy numbers are generated from the forecast data, covering forecast data and actual data. Again using a distance function, actual data is compared with the fuzzy number to determine how the grade of membership is. The model was prepared in such a way that all the data which is usually imprecise, chaotic, uncertain can be used. The model is used in Poland by the Institute of Meteorology and by Water Management, and by the Voivodship Inspector for Environmental Protection. It forecast selected pollution concentrations for all areas of Poland.     

基于大数据模式识别机器学习算法的热力站动态能耗指标预测模型

为了实现对供热系统热力站的热负荷预测,将天气、用户室内温度和时间迟滞性等因素作为负荷预测的数据依据。本文利用统计学原理和大数据架构,通过大量数据样本的学习和修正,为解决热力行业热力控制理论缺失问题提供了一条崭新的方法。将模式识别算法和时间序列相关性分析作为算法的核心,为解决天气和时间迟滞性对用户供热的影响提供了可能。本文以我公司n个典型热力站和其所带热用户为实验对象,以所在地区天气预报和天气实时数据为依据,对供暖期间所采集的热力站供暖数据、天气预报数据、典型供暖用户室内温度,通过大数据模式识别机器学习算法对样本进行学习训练,形成一套完整热力站动态能耗指标预测模型。     

基于 HY-FMV 模型的雷暴天气预测研究

本文以极端天气中的雷暴天气为研究对象,基于历史气象数据预测未来三小时是否发生雷暴。为预测雷暴是否发生,本文分别对极端天气气象数据的采样、数据预处理、特征选择,以及建模分析进行了研究,最终提出一种基于机器学习方法的HY-FMV模型框架对雷暴天气进行预测。该模型采用混合模型进行数据预处理,基于概率分布与模型评价进行特征的选择和构建,并使用梯度提升树算法对极端天气进行预测分类。最后,本文以2010年到2015年福建和广东两省数据为例,分别使用本文所提出的HY-FMV模型,和随机森林算法等进行雷暴天气预测,结果表明,本文所提出的HY-FMV模型在F1指标上精度达到78%,相比其他算法,在雷暴天气预测精度上提高了0.5%-0.6%。     

On-line economic optimization of energy systems using weather forecast information

We establish an on-line optimization framework to exploit weather forecast information in the operation of energy systems. We argue that anticipating the weather conditions can lead to more proactive and cost-effective operations. The framework is based on the solution of a stochastic dynamic real-time optimization (D-RTO) problem incorporating forecasts generated from a state-of-the-art weather prediction model. The necessary uncertainty information is extracted from the weather model using an ensemble approach. The accuracy of the forecast trends and uncertainty bounds are validated using real meteorological data. We present a numerical simulation study in a building system to demonstrate the developments.     

Hybrid PSO–SVM method for short-term load forecasting during periods with significant temperature variations in city of Burbank

This paper proposes a practical new hybrid model for short term electrical load forecasting based on particle swarm optimization (PSO) and support vector machines (SVM). Proposed PSO–SVM model is targeted for forecast load during periods with significant temperature variations. The proposed model detects periods when temperature significantly changes based on weather (temperature) forecast and decides whether the model can be trained just on recent history (typically 4 weeks ago) or such history has to be modified with data for similar days taken from history beyond recent history when such weather conditions were detected. Architecture of the solution consists of three modules, preprocessing module, SVM module and PSO module. The algorithm has been tested in city of Burbank utility, USA and obtained results show better accuracy comparing to results generated with classical methods of training on recent history only or similar days only.     

Assessment and weighting of meteorological ensemble forecast members based on supervised machine learning with application to runoff simulations and flood warning

Numerical weather forecasts, such as meteorological forecasts of precipitation, are inherently uncertain. These uncertainties depend on model physics as well as initial and boundary conditions. Since precipitation forecasts form the input into hydrological models, the uncertainties of the precipitation forecasts result in uncertainties of flood forecasts. In order to consider these uncertainties, ensemble prediction systems are applied. These systems consist of several members simulated by different models or using a single model under varying initial and boundary conditions. However, a too wide uncertainty range obtained as a result of taking into account members with poor prediction skills may lead to underestimation or exaggeration of the risk of hazardous events. Therefore, the uncertainty range of model-based flood forecasts derived from the meteorological ensembles has to be restricted.In this paper, a methodology towards improving flood forecasts by weighting ensemble members according to their skills is presented. The skill of each ensemble member is evaluated by comparing the results of forecasts corresponding to this member with observed values in the past. Since numerous forecasts are required in order to reliably evaluate the skill, the evaluation procedure is time-consuming and tedious. Moreover, the evaluation is highly subjective, because an expert who performs it makes his decision based on his implicit knowledge.Therefore, approaches for the automated evaluation of such forecasts are required. Here, we present a semi-automated approach for the assessment of precipitation forecast ensemble members. The approach is based on supervised machine learning and was tested on ensemble precipitation forecasts for the area of the Mulde river basin in Germany. Based on the evaluation results of the specific ensemble members, weights corresponding to their forecast skill were calculated. These weights were then successfully used to reduce the uncertainties within rainfall-runoff simulations and flood risk predictions.