Hybrid Variational Mode Decomposition and evolutionary robust kernel extreme learning machine for stock price and movement prediction on daily basis

The Empirical Mode Decomposition (EMD) has been applied successfully in many forecasting problems. The Variational Mode Decomposition (VMD), a more effective decomposition technique has been proposed with an aim to avoid the limitations of EMD. This study focuses on two objectives i.e. day ahead stock price prediction and daily trend prediction using Robust Kernel based Extreme Learning Machine (RKELM) integrated with VMD where the kernel function parameters optimized with Differential Evolution (DE) algorithm here named as DE-VMD-RKELM. These experiments have been conducted on BSE S&P 500 Index (BSE), Hang Seng Index (HSI) and Financial Times Stock Exchange 100 Index (FTSE), and the daily price prediction performance of the proposed VMD-RKELM model is measured in terms of Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE). On the other hand the daily trend prediction which is defined as a classification problem is measured in terms of Percentage of Correct Classification Accuracy (PCCA). The prediction performance of the VMD-RKELM is compared with the performance of robust Extreme Learning Machine (RELM), Extreme Learning Machine integrated with EMD (EMD-RELM). Robust Kernel Extreme Learning Machine integrated with EMD (EMD-RKELM) and two benchmark approaches i.e. Support Vector Regression (SVR) and Autoregressive Moving Average (ARMA). The trend prediction results are compared with Naive–Bayes classifier, ANN (artificial neural network), and SVM (support vector machine). The experimental results obtained from this study for price prediction as well as trend classification performance are promising and the prediction analysis illustrated in this work proves the superiority of the VMD-RKELM model over the other predictive methods.     

粗纱工艺参数相关分析及BP网络预报研究

在标准化企业粗纱工序生产数据的基础上,针对神经网络输入端参数组会影响最终预报结果的特点,提出分别利用相关性分析法和多元逐步回归分析法筛选对粗纱CV值(R1)和单重(R2)影响较大的参数。将筛选出的参数按重要程度由大到小依次输入BP网络,采用多输入单输出子网组方式建立了4个网络模型。训练好的模型经10组检验样本检验,其预报结果和实测结果的平均相对误差(MEP)都低于4%。用20组未参与建模的验证数据进行预报表明:相关性分析法筛选参数建立的模型对R1和R2的绝对值平均预报精度分别为2.63%和2.98%,且预报值与实测值间的相关系数分别为0.884和0.958,这些指标都优于采用多元逐步回归分析法筛选参数建立的模型。     

Demand forecasting of perishable farm products using support vector machine

This article presents a new algorithm for forecasting demand for perishable farm products, based on the support vector machine (SVM) method. Since SVMs have greater generalisation performance and guarantee global minima for given training data, it is believed that support vector regression will perform well for forecasting demand for perishable farm products. In order to improve forecasting precision (FP), this article quantifies the factors affecting the sales forecast of perishable farm products based on the fuzzy theory, which is suitable for real situations. Numerical experiments show that forecasting systems with SVMs and fuzzy theory outperform the radial basis function neural network, based on the criteria of day absolute error, relative mean error and FP. Since there is no structured way to choose the free parameters of SVMs, the variational range of free parameters and the effects of the parameters on prediction performance are discussed in this article. Analysis of experimental results proves that it is advantageous to apply SVMs forecasting system in perishable farm products demand forecasting.     

相关向量机超参数优化的网络安全态势预测

针对当前网络安全态势预测方法存在的过学习与欠学习、自由参数多、预测精度不高等问题,提出使用一种改进模拟退火法优化的相关向量机模型(PSA-RVM)来解决网络安全态势预测问题。在预测过程中,首先对网络安全态势样本数据进行相空间重构形成训练样本集;然后,利用Powell算法改进模拟退火(PSA)法,并将相关向量机(RVM)嵌入到PSA算法的目标函数计算过程中,优化RVM超参数,以得到学习能力、预测精度提升的网络安全态势预测模型。仿真实例表明,所提方法具有较高的预测精度,平均相对误差(MAPE)和均方根误差(RMSE)分别为0.39256和0.01261,均优于Elman和PSO-SVR模型;所提方法能够较好地刻画网络安全态势的变化趋势,有助于网络管理人员把握未来网络安全态势发展趋势,从而提前主动采取相应的网络防御措施。     

ε-Descending Support Vector Machines for Financial Time Series Forecasting

This paper proposes a modified version of support vector machines (SVMs), called ε-descending support vector machines (ε-DSVMs), to model non-stationary financial time series. The ε-DSVMs are obtained by incorporating the problem domain knowledge – non-stationarity of financial time series into SVMs. Unlike the standard SVMs which use a constant tube in all the training data points, the ε-DSVMs use an adaptive tube to deal with the structure changes in the data. The experiment shows that the ε-DSVMs generalize better than the standard SVMs in forecasting non-stationary financial time series. Another advantage of this modification is that the ε-DSVMs converge to fewer support vectors, resulting in a sparser representation of the solution. This revised version was published online in June 2006 with corrections to the Cover Date.     

基于Elman神经网络的GNSS/INS全域高精度定位方法

针对当前智能网联汽车定位与导航系统无法接收全球导航卫星系统（GNSS）信号引起定位失效的问题，提出一种基于Elman神经网络的GNSS结合惯性导航系统（INS）的全域高精度定位方法。首先，采用神经网络方法，建立了基于Elman网络的GNSS/INS高精度定位训练模型和GNSS失效预测模型；然后，利用GNSS、INS和实时动态（RTK）等定位技术，设计了GNSS/INS高精度定位数据采集实验系统；最后，选取采集的有效实验数据进行了反向传播（BP）神经网络、级联BP（CFBP）神经网络、Elman神经网络的训练模型性能对比分析，并验证了基于Elman网络的GNSS失效预测模型。实验结果表明，所提方法训练误差指标均优于基于BP和CFBP神经网络的方法；在GNSS失效1 min、2 min、5 min时，基于预测模型的预测平均绝对误差（MAE）、方差（VAR）和均方根误差（RMSE）分别为18.88 cm、19.29 cm、58.83 cm，8.96、8.45、5.68和20.90、21.06、59.10，随着GNSS信号失效时长的增加，定位预测精度降低。     

基于加权K-Means和局部BPNN的票房预测模型

电影作为典型的短周期、体验型产品,其票房收益受众多因素的共同影响,因此对其票房进行预测较为困难.本文主要构建了一种基于加权K-均值以及局部BP神经网络（BPNN）的票房预测模型对目前的票房预测模型存在的不足进行改进,从而提高票房预测的精度：（1）构建基于随机森林的影响因素影响力测量模型,并以此为依据对票房影响因素进行筛选,以此来简化后续预测模型的输入;（2）考虑到不同影响因素对票房的影响力不同的现实情况,为了解决以往研究中对影响因素权重平均分配的问题,本文构建了基于加权K-均值和局部BP神经网络的票房预测模型,以因素影响力为依据对样本数据进行加权的K-均值聚类,并基于子样本构建局部BP神经网络模型进行票房预测.实验证明,本文所构建的模型平均绝对百分比误差（MAPE）为8.49%,低于对比实验的10.39%,可以看出本文构建的基于加权K-均值以及局部BP神经网络的票房预测模型的预测结果要优于对比模型的预测结果.     

基于贝叶斯模型的云服务服务质量预测

针对如何分配一个未来一段时间内满足QoS要求的云服务和感知可能将要发生的QoS违规的问题,提出一种基于时间序列预测方法的云服务QoS预测方法。该预测方法利用改进的贝叶斯常均值(IBCM)模型,能够准确地预测云服务未来一段时间内的QoS状态。实验通过搭建Hadoop集群模拟云平台并收集了响应时间和吞吐量两种QoS属性的数据作为预测对象,实验结果表明:相比自回归积分滑动平均(ARIMA)模型和贝叶斯常均值折扣模型等时间序列预测方法,基于改进的贝叶斯常均值模型的云服务QoS预测方法的平方和误差(SSE)、平均绝对误差(MAE)、均方误差(MSE)和和平均绝对百分比误差(MAPE)均比前两者小一个数量级,因此具有更高的预测精度;同时预测结果对比图说明提出的预测方法具有更好的拟合效果。     

嵌入互联网舆情强度的人民币汇率预测

针对目前人民币汇率预测研究存在的数据源单一导致难以提升预测效果的问题,提出一种嵌入互联网舆情强度的预测技术,通过融合多方面数据源进行对比分析,有效降低了人民币汇率的预测误差。首先,融合互联网外汇新闻数据和历史行情数据,并将多源文本数据转化为可计算的特征向量;其次,通过情感特征向量构建五种特征组合并对其进行对比,给出了嵌入互联网舆情强度的特征组合作为预测模型输入;最后,设计外汇舆情影响汇率预测的滑动时间窗口,建立基于机器学习的汇率预测模型。实验结果表明,嵌入互联网舆情的特征组合相对于不含舆情的特征组合在均方根误差（RMSE）和平均绝对误差（MAE）上分别提升了9.8%和16.2%;此外,长短期记忆网络（LSTM）预测模型比支持向量回归（SVR）、决策回归（DT）和深度神经网络（DNN）预测模型表现更好。     

Dynamic Ensemble Multivariate Time Series Forecasting Model for PM2.5

In forecasting real time environmental factors, large data is needed to analyse the pattern behind the data values. Air pollution is a major threat towards developing countries and it is proliferating every year. Many methods in time series prediction and deep learning models to estimate the severity of air pollution. Each independent variable contributing towards pollution is necessary to analyse the trend behind the air pollution in that particular locality. This approach selects multivariate time series and coalesce a real time updatable autoregressive model to forecast Particulate matter (PM) PM2.5. To perform experimental analysis the data from the Central Pollution Control Board (CPCB) is used. Prediction is carried out for Chennai with seven locations and estimated PM’s using the weighted ensemble method. Proposed method for air pollution prediction unveiled effective and moored performance in long term prediction. Dynamic budge with high weighted k-models are used simultaneously and devising an ensemble helps to achieve stable forecasting. Computational time of ensemble decreases with parallel processing in each sub model. Weighted ensemble model shows high performance in long term prediction when compared to the traditional time series models like Vector Auto-Regression (VAR), Autoregressive Integrated with Moving Average (ARIMA), Autoregressive Moving Average with Extended terms (ARMEX). Evaluation metrics like Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and the time to achieve the time series are compared.