基于加权组合模型的红外时间序列预测

为克服Prophet模型对残差自相关性考虑的缺失,时间推理能力的不足,提高被动红外(passive infrared,PIR)运动探测器检测结果的准确性,提出一种Prophet与SARIMA动态加权组合的预测模型.分析PIR运动探测器的特点,分析对比几种流行的预测算法的优劣,在此基础上设计Prophet-SARIMA的组合预测模型,统计和分析用户的行为.为获取最好的组合效果,设计动态加权组合算法,通过加权算法可确定最优的权值组合.通过实验验证了Prophet-SARIM A组合预测模型在人体红外数据的预测中具有更强的适用性和更高的准确性.     

提升基于网络的大数据预测分析能力的方法

大数据时代,多样化的海量数据,可以利用互联网技术和云计算技术提供科学有效的数据存储技术、数据挖掘算法和智能的语义引擎,实现高质量的数据挖掘、管理和调用,提供更强的洞察力、决策力和精准的预测分析能力。基于此,笔者提出提升基于网络的大数据预测分析能力的方法,以供参考。     

融合拓扑和属性的动态网络链路预测方法

网络数据中出现的大量节点属性和随时间变化的特征,给链路预测提出了新挑战。基于注意力机制和循环神经网络对随时间演化网络进行建模,提出了DTA-LP模型。与传统的静态链路预测算法相比,DTA-LP使用LSTM捕获时序信息,动态预测可以更好应用于现实网络;与基于网络拓扑的动态链路预测算法相比,DTA-LP可以聚集高阶拓扑特征,有效挖掘网络邻域信息;与基于属性网络的动态链路预测算法相比,DTA-LP可以加权融合网络拓扑属性,提高预测精度。在4种真实数据上的实验结果表明,该方法能结合网络已有先验知识,以较高的MAP值来预测未来网络中的边,验证了模型的有效性。     

基于多嵌入维数的风力发电功率组合预测模型

为了减小混沌系统的重构参数对预测结果的影响,提出了基于多嵌入维数的风力发电功率组合预测模型.分别使用线性加权算法和神经网络算法对单一的基于相空间重构的神经网络模型进行组合,既综合了各嵌入维数下的信息,又将各维数下的预测偏差进行融合,从而有效提高了预测精度.通过对黑龙江富锦风电场的功率时间序列进行验证,证实了该组合模型的有效性,神经网络非线性组合算法的预测误差小于7%.     

改进自适应学习树电源管理预测策略

为了降低嵌入式设备的功耗,研究了基于自适应学习树结构模型的动态电源管理预测策略.通过在基于概率自适应学习树结构模型的基础上添加空闲时间长度结点,提出了概率统计加权空闲时间的改进自适应学习树电源管理预测策略,以空闲时间长度作为预测依据,同时采用实际状态历史概率统计的结果进行预测空闲时间长度的更新.仿真结果表明,该方法可以有效地降低设备功耗,并且提高了预测准确率.     

基于XGBoost和LSTM加权组合模型在销售预测的应用

针对多变量的商品销售预测问题,为了提高预测的精度,提出了一种ARIMA-XGBoost-LSTM加权组合方法,对具有多个影响因素的商品销售序列进行预测,本文采用ARIMA做单变量预测,将预测值作为新变量同其他变量一起放入XGBoost模型中进行不同属性的挖掘,并将XGBoost的预测值合并到多变量序列中,然后通过将新的多维数据转换为监督学习序列后利用LSTM模型进行预测,将3种模型预测结果进行加权组合,通过多次实验得出最佳组合的权值,以此计算出最终的预测值.数据结果表明,基于XGBoost和LSTM的加权组合的多变量预测方法比单一的预测方法所得到的预测值更为精准.     

基于动态加权集成学习的遥测数据预测方法

飞行任务中的遥测数据是飞行器中各功能子系统监测模块顺序产生的多维时间序列,其反应各子系统功能是否正常,对遥测数据的精准预测是研判飞行器运行状态的重要依据;针对已有时间序列预测算法会随时间劣化的缺点,提出基于集成学习原理的动态加权神经网络集成算法;该方法通过神经网络强数据拟合能力,集成学习算法具有的泛化特性和动态加权算法适应数据的漂移变化特性,提升算法的整体预测精度;选择多层感知机神经网络作为基学习器,给出神经网络基学习器结构设计方法和优化方法,以及动态加权算法的具体过程,将其与静态加权算法进行比较实验,该算法对预测精度提高效果显著,一定程度抑制数据的漂移,结果表明采用动态加权集成学习适合对遥测数据的预测.     

智慧农业中时序数据组合预测模型

智慧农业是实现农业精准化的技术解决方案,智慧农业系统可以实时监测植物生长的各类环境参数,并可以应用相应的预测模型来模拟农作物生长环境的变化趋势,为科学决策提供依据。近年来有很多学者提出了时间序列的预测模型算法,在预测稳定性方面取得了不错的效果。为了进一步提升时间序列的预测精度,提出一种基于差分整合移动平均自回归模型和小波神经网络的组合预测模型。该组合模型结合2个单项模型优点,用差分整合移动平均自回归模型来拟合序列的线性部分,用小波神经网络来校正其残差,使其拟合曲线更接近于实际值,采用温室内的历史温度数据来验证该组合模型的精确度,最后将组合模型与传统预测模型的预测结果进行对比。结果表明,该组合模型用于温室温度预测的精确度更高,拟合效果更好,相比于传统模型预测算法计算效能提高了20%左右。     

基于CatBoost用信预测模型的TreeSHAP解释性研究

银行客户申请信用贷款在授信通过后,精准预测客户是否用信及分析影响客户用信的关键因素,对提高银行客户服务能力及盈利能力具有重要意义.目前,机器学习算法鲜有在用信预测方面的应用,且金融用信领域缺乏模型可解释性的研究,为此提出一种基于CatBoost的TreeSHAP解释性用信预测模型.通过CatBoost构建用信预测模型,利用3种超参数优化算法对该模型进行对比优化,与基线模型在4项主要性能指标上进行实验对比,结果表明经TPE算法优化后的模型性能均优于其他模型,然后结合TreeSHAP方法从全局和局部的层面增强模型的可解释性,解释性分析客户用信的影响因素,为银行对客户进行精准化营销提供决策依据.     

基于5G与CNN的智能电网稳定性预测

随着5G通信技术的研究以及新型基础设施的建设, 智能电网得到了快速发展. 同时, 在大数据时代, 万物互联导致海量的设备接入电力网络, 也给智能电网带来了较大的负担, 电力网络的稳定性问题亟待解决. 因此, 本文提出了一种基于CNN的智能电网稳定性预测算法, 通过收集电力网络产生的数据, 经过CNN模型的处理, 最后输出智能电网稳定性的判别结果. 经过仿真验证, 该算法与SVM、AdaBoost, 随机森林相比, 具有较高的准确率; 同时, 本文采用了4种不同的优化算法去改善CNN模型, 带有动量的SGD算法可以达到98.13％预测准确度, 利用该模型可以有效帮助电力系统对未知的问题提前预警, 降低了安全隐患并减少了电力事故的发生.     

航材消耗广义加权函数比例平均组合预测模型

提出一种新的组合预测模型--广义加权函数比例平均组合预测模型,并利用二次规 划算法给出其加权系数的参数估计方法.同时,针对航材消耗的季节性与波动性特点,建立了航 材消耗预测的灰色系统模型与神经网络模型,最后建立了基于灰色系统与神经网络的航材消耗 广义加权函数比例平均组合预测模型并以实例说明了其预测效果.     

基于特征选择和iJaya-SVR的年度电力消费预测研究

准确的电力消费预测对能源规划和政策制定具有重要意义.鉴于已有研究忽略了特征冗余以及智能优化算法控制参数不确定对预测精度的影响,引入最大相关最小冗余(MRMR)算法筛选电力消费的关键影响因素作为预测指标,提出改进的Jaya算法(iJaya)用于优化支持向量回归(SVR)的超参数,进而构建MRMR-iJaya-SVR预测模型.以我国的年度电力消费数据为例,对MRMR-iJaya-SVR模型的预测效果进行验证,并利用北京市的年度电力消费数据测试其鲁棒性.结果表明:iJaya算法具有较强的全局搜索能力和较好的稳定性,MRMR-iJaya- SVR模型在单步预测和多步预测中的表现均优于基准模型.此外,对于不同的数据集,MRMR-iJaya-SVR模型均具有良好的鲁棒性.     

基于IPSO-FHMM的非侵入式负荷分解

非侵入式负荷分解是智能用电系统的一个重要环节,可深入分析用户的用电信息,对负荷预测、需求侧管理及电网安全有重要意义.本文提出了一种基于改进粒子群优化因子隐马尔可夫模型(IPSO-FHMM)的非侵入式负荷分解方法.利用高斯混合模型(GMM)对单负荷进行状态聚类,总负载模型由因子隐马尔可夫模型表示.针对Baum-Welch算法容易收敛于局部极值的问题,将线性递减权重的粒子群优化算法引入到FHMM的参数训练中.使用AMPds2数据集进行仿真实验,结果表明,该模型可以有效地提高分解精度.     

改进IOWHA算子组合预测模型

针对现有单项预测模型提供信息有限,预测误差大的问题,引用最优加权组合建模理论,将灰色关联度与IOWHA算子相结合,提出一种新的组合预测模型权重确定方法,并应用该权重确定方法构建了一种基于RBF神经网络预测模型和GM预测模型的最优组合预测模型。该模型能够克服传统组合预测方法的两个缺陷：加权平均系数不变和以单一误差指标为准则。利用该组合模型对全国物流需求进行组合预测,并与RBF神经网络模型、GM模型的预测结果进行了对比分析。结果表明,相对于单项预测模型,该组合预测模型的预测精度更高,是一种有效的物流需求预测模型。     

Intelligent integrated data processing model for oceanic warning system

There are a number of dirty data in observation data set derived from ocean observing network. These data should be carefully and reasonably processed before they are used for forecasting or analysis in oceanic warning system (OWS). Due to high-dimensional and dynamic oceanic data, we propose an intelligent integrated data processing model for the OWS. Firstly, we design an integrated framework of the oceanic data processing and present its processing model. The function of each module of this model is analyzed in details. Then, we propose several intelligent data processing methods, such as an intelligent data cleaning method based on the fuzzy c-means algorithm, a data filtering and clustering method based on the greedy clustering algorithm, and a data processing method based on the maximum entropy for the OWS. The efficiency and accuracy of the proposed model is proved by experimental results of observation data of the Red Tide. The proposed model can automatically find the new clustering center with the updated sample data, and outperforms several algorithms in data processing for the OWS.     

Variable weights combined model based on multi-objective optimization for short-term wind speed forecasting

Accurate and steady wind speed prediction is essential for the efficient management of wind power factories and energy systems. However, it is difficult to obtain satisfactory forecasting performance because of the characteristics of random nonlinear fluctuations inherent in wind speed variation. Considering the drawbacks of statistical models in forecasting nonlinear time series and the problem of artificial intelligence models easily falling into a local optimum, in this study, we successfully integrate the variable weighted combination theory into a new combined forecasting model that simultaneously consists of three disparate hybrid models based on the decomposition technology. Moreover, the extreme learning machine optimized by the multi-objective grasshopper optimization algorithm is adopted to integrate all the forecasting results derived from each hybrid model to further enhance the forecasting accuracy. In this study, we consider a case study that employs several authentic wind speed data aggregates of Shandong wind farms for an evaluation of the forecasting performance of the proposed combined model. The experimental results reveal that this proposed model surpasses the contrasted benchmark models and is satisfactory for intellective grid programs.     

基于最优定权组合法的大气污染物SO2预测

为了提高对大气污染物SO₂的预测准确率,基于多个空气质量预测模式（WRF-CHEM、CMAQ、CAMx）,以过去一段时间内各单项空气质量预测模式的组合预测误差平方和最小为原则,构建出针对大气污染物SO₂的最优定权组合预测模型.选取2018年云南省楚雄、昭通、蒙自三个站点1至5月份的实际观测数据和前述三个空气质量模式的预测数据作为实验样本,然后分别采用多元线性回归法和动态权重更新法在相同的实验条件下与所提的最优定权组合预测法进行预测对比实验.实验结果表明,所提方法的预测值相较其他两种方法更加贴近实际观测值,其两项误差评估指标值均最小.总体而言,最优定权组合预测模型很好地综合了各单项空气质量预测模式的优势,提高了SO₂的预测精度.     

ECA rule learning in dynamic environments

Through the development of management and intelligent control systems, we can make useful decision by using incoming data. These systems are used commonly in dynamic environments that some of which are been rule-based architectures. Event–Condition–Action (ECA) rule is one of the types that are used in dynamic environments. ECA rules have been designed for the systems that need automatic response to certain conditions or events. Changes of environmental conditions during the time are important factors impacting a reduction of the effectiveness of these rules which are implied by changing users demands of the systems that vary over time. Also, the rate of the changes in the rules are not known which means we are faced with the lack of information about rate of occurrence of new unknown conditions as a result of dynamics environments. Therefore, an intelligent rule learning is required for ECA rules to maintain the efficiency of the system. To the best knowledge of the authors, ECA rule learning has not been investigated. An intelligent rule learning for ECA rules are studied in this paper and a method is presented by using a combination of multi flexible fuzzy tree (MFlexDT) algorithm and neural network. Hence data loss could be avoided by considering the uncertainty aspect. Owing to runtime, speed, and also stream data in dynamic environments, a hierarchical learning model is proposed. We evaluate the performance of the proposed method for resource management in the Grid and e-commerce as case studies by modeling and simulating. A case study is presented to show the applicability of the proposed method.     

An improved whale optimization algorithm for forecasting water resources demand

Water demand forecasting can promote the rational use of water resources and alleviate the pressure on water demand. By analyzing the use of water resources, this paper establishes three models of water demand forecasting, logarithmic model, linear and exponential combination model and linear, exponential and logarithmic hybrid models. In order to accurately estimate the demand for water resources, an improved whale optimization algorithm based on social learning and wavelet mutation strategy is proposed. The new algorithm designs a new linear incremental probability, which increases the possibility of global search of the algorithm. Based on the social learning principle, the social ranking and social influence are used to construct the social network for the individual, and the adaptive neighborhood learning strategy based on the network relationship is established to achieve the exchange and sharing of information between groups. The Morlet wavelet mutation mechanism is integrated to realize the dynamic adjustment of the mutation space, which enhances the ability of the algorithm to escape from local optimization. The latest CEC2017 benchmark functions confirms the superiority of the proposed algorithm. The water consumption from 2004 to 2016 in Shaanxi Province of China is used for the experiment. The results show that the performance of the proposed algorithm for solving the three water resources forecasting model is better in comparison to other algorithms. The prediction accuracy is as high as 99.68%, which verified the validity of the model and the practicality of the proposed algorithm.