灰色正交化方法在用电量预测中的仿真研究

根据灰色正交化方法和马尔可夫链原理,应用Gauss-Chebyshev正交化思想预测时序数据的总体趋势。预测的精度是时变的,而马尔可夫链原理在处理时变的系统过程时具有较好的优势,选用该方法能更好的解决预测结果的不稳定性。基于此,提出一种用于用电量数据预测的灰色马尔可夫正交化模型,适用于中短期、数据需求量少且数据振幅较大的动态过程预测。最后用提出的方法对江苏省2007年工业用电量进行预测,其结果表明了所提方法的有效性。

Abstract：

The general trend of time series data was predicted with Gauss-Chebyshev orthogonalization theory according to the grey orthogonalization method and the Markov Chain theory.The prediction accuracy is time-varying.However,this approach will better solve the problem of unstable prediction result since Markov chain theory has greater advantages in handling time-varying system process.Based on this,the Markov grey orthogonalization model prediction was proposed for electricity consumption.It is suitable for dynamic process prediction in medium and short term with less data demand and large data fluctuations.Finally,the proposed approach was used to forecast the industrial electricity consumption of Jiangsu Province in 2007,and the results show the effectiveness of this approach.     

基于抗差最小二乘配置的海底地形生成研究

最小二乘配置算法能同时顾及测深数据的系统性和随机性影响,提高海底地形的生成精度。根据该算法建立了海洋测深多波束数据的函数模型和随机模型,并构建海底地形。为避免测深数据的异常值影响,进一步推导了基于抗差估计的协方差函数求解公式,提出抗差最小二乘配置算法生成海底地形。利用多波束数据生成海底,结果显示抗差最小二乘配置算法能在构建海底地形的同时准确剔除测深异常。将构建的地形与双线性多项式内插生成的海底进行比较,进一步说明了该算法具有较高的海底地形生成精度。

Abstract：

Seabed terrain generating precision can be improved by least-squares collocation algorithm which takes the systematic and stochastic effects in the bathymetry data. The functional model and stochastic model of the algorithm were created by multibeam bathymetry data,and the method of creating seabed terrain by Least-squares Collocation was researched. To avoid the effect of outliers in bathymetry data,covariance function was calculated by robust estimation,and the Robust Least-squares Collocation algorithm for terrain generation was proposed. It was applied to the real bathymetry data set,and the results indicate that the outliers are detected by the algorithm while the seabed terrain is generated. In the remainder,the terrain grids was compared with which created by bilinear polynomial interpolation algorithm,and it is proved that the Robust Least-squares Collocation algorithm can get higher precision of seabed terrain generation while detecting outliers.     

Parameter estimation for dual-rate sampled Hammerstein systems with dead-zone nonlinearity

The identification of nonlinear systems with multiple sampled rates is a difficult task.The motivation of our paper is to study the parameter estimation problem of Hammerstein systems with dead-zone characteristics by using the dual-rate sampled data.Firstly,the auxiliary model identification principle is used to estimate the unmeasurable variables,and the recursive estimation algorithm is proposed to identify the parameters of the static nonlinear model with the dead-zone function and the parameters of the dynamic linear system model.Then,the convergence of the proposed identification algorithm is analyzed by using the martingale convergence theorem.It is proved theoretically that the estimated parameters can converge to the real values under the condition of continuous excitation.Finally,the validity of the proposed algorithm is proved by the identification of the dual-rate sampled nonlinear systems.     

Multi-label dimensionality reduction and classification with extreme learning machines

In the need of some real applications, such as text categorization and image classification, the multi-label learning gradually becomes a hot research point in recent years. Much attention has been paid to the research of multi-label classification algorithms. Considering the fact that the high dimensionality of the multi-label datasets may cause the curse of dimensionality and wil hamper the classification process, a dimensionality reduction algorithm, named multi-label kernel discriminant analysis （MLKDA）, is proposed to reduce the dimensionality of multi-label datasets. MLKDA, with the kernel trick, processes the multi-label integrally and realizes the nonlinear dimensionality reduction with the idea similar with linear discriminant analysis （LDA）. In the classification process of multi-label data, the extreme learning machine （ELM） is an efficient algorithm in the premise of good accuracy. MLKDA, combined with ELM, shows a good performance in multi-label learning experiments with several datasets. The experiments on both static data and data stream show that MLKDA outperforms multi-label dimensionality reduction via dependence maximization （MDDM） and multi-label linear discriminant analysis （MLDA） in cases of balanced datasets and stronger correlation between tags, and ELM is also a good choice for multi-label classification.     

Model algorithm control using neural networks for input delayed nonlinear control system

The performance of the model algorithm control method is partially based on the accuracy of the system’s model. It is difficult to obtain a good model of a nonlinear system, especially when the nonlinearity is high. Neural networks have the ability to "learn"the characteristics of a system through nonlinear mapping to represent nonlinear functions as well as their inverse functions. This paper presents a model algorithm control method using neural networks for nonlinear time delay systems. Two neural networks are used in the control scheme. One neural network is trained as the model of the nonlinear time delay system, and the other one produces the control inputs. The neural networks are combined with the model algorithm control method to control the nonlinear time delay systems. Three examples are used to illustrate the proposed control method. The simulation results show that the proposed control method has a good control performance for nonlinear time delay systems.     

Parameter selection of support vector machine for function approximation based on chaos optimization

The support vector machine （SVM） is a novel machine learning method, which has the ability to approximate nonlinear functions with arbitrary accuracy. Setting parameters well is very crucial for SVM learning results and generalization ability, and now there is no systematic, general method for parameter selection. In this article, the SVM parameter selection for function approximation is regarded as a compound optimization problem and a mutative scale chaos optimization algorithm is employed to search for optimal paraxneter values. The chaos optimization algorithm is an effective way for global optimal and the mutative scale chaos algorithm could improve the search efficiency and accuracy. Several simulation examples show the sensitivity of the SVM parameters and demonstrate the superiority of this proposed method for nonlinear function approximation.     

Study on the Hungarian algorithm for the maximum likelihood data association problem

A specialized Hungarian algorithm was developed here for the maximum likelihood data association problem with two implementation versions due to presence of false alarms and missed detections. The maximum likelihood data association problem is formulated as a bipartite weighted matching problem. Its duality and the optimality conditions are given. The Hungarian algorithm with its computational steps, data structure and computational complexity is presented. The two implementation versions, Hungarian forest （HF） algorithm and Hungarian tree （HT） algorithm, and their combination with the naYve auction initialization are discussed. The computational results show that HT algorithm is slightly faster than HF algorithm and they are both superior to the classic Munkres algorithm.     

Method of Chaotic Prediction Based on Wavelet BP Network and Its Application

The advantage of artificial neural network and wavelet analysis are integrated through replacing the traditional S-shaped activation function with the wavelet function. One method of chaotic prediction based on wavelet BP network was put forward based on the reconstruction of state space. Training data construction and networks structure are determined by chaotic phase space, and nonlinear relationship of phase points was established by BP neural networks. As an example, the new method was applied on short term forecasting of monthly precipitation time series of Sanjiang Plain with chaotic characteristics. The results showed so higher precision of the method had that the theoretical evidence would be provided for applying the chaos theory to study the variable law of monthly precipitation.     

DBS imaging based on Keystone transform

To correct the range walk through resolution cell in Doppler beam sharpening(DBS) imaging,a new DBS imaging algorithm based on Keystone transform is proposed.Without the exact values of the movement parameters and the look angle of the radar platform in the multi-targets environment,a linear transform on the received data is employed to correct different range walk values accurately under the condition of Doppler frequency ambiguity in this algorithm.This method can realize the coherent integration in azimuth dimension and improve the azimuth resolution.In order to reduce the computational burden,a fast implementation of Keystone transform is used.Theoretical analysis and simulation results demonstrate the effectiveness of the new algorithm.And through comparing the computational load of the fast implementation with several other algorithms,the real-time processing ability of the proposed algorithm is superior to that of other algorithms.     

ERBF network with immune clustering

Based on immune clustering and evolutionary programming(EP), a hybrid algorithm to train the RBF network is proposed. An immune fuzzy C-means clustering algorithm (IFCM) is used to adaptively specify the amount and initial positions of the RBF centers according to input data set; then the RBF network is trained with EP that tends to global optima. The application of the hybrid algorithm in multiuser detection problem demonstrates that the RBF network trained with the algorithm has simple network structure with good generalization ability.     

峰度检验加窗的EMD脉冲星信号消噪方法

针对经验模态分解(empirical mode decomposition, EMD)过程中本征模态函数(intrinsic mode function, IMF)上存在脉冲星信号与噪声混叠的问题,提出了一种基于局部峰度检验加窗的EMD消噪方法。首先,利用自相关和互相关来判断IMF的重构起点;其次,通过局部峰度检验方法来获取重构起点前两层IMF中信号脉冲部分的左、右端点,选用Turkey Hanning窗滤掉脉冲外噪声;最后,利用自适应阈值方法进一步除噪,改善信号质量。实验结果表明,与其他5种方法相比,所提消噪方法可以有效抑制噪声,保留脉冲星信号细节信息,具有更高的消噪性能。     

基于EMD能量占比的海面漂浮小目标特征检测

针对传统的时频分析方法对海杂波分析有限的问题,提出一种基于经验模态分解(empirical mode decomposition,EMD)能量占比的海面漂浮小目标特征检测方法.首先,采用EMD将接收回波分为独立不同尺度的若干个固有模态(intrinsic mode function,IMF)分量,实现对接收回波的频率从...     

消除EMD端点效应的PSO-SVM方法研究

经验模态分解(empirical mode decomposition, 简称EMD)的端点效应使得EMD分解结果产生严重失真, 为了减小分解过程中产生的端点效应, 将支持向量机(SVM)这一智能算法引入EMD, 提出采用SVM模型解决分解中产生的端点效应问题. 通过支持向量机对其原始数据两端进行延拓, 以获得一个或者多个极大值和极小值. 为了使端点处的延拓变得更加合理, 引入粒子群(PSO)智能算法对支持向量机算法参数进行优化, 使其两个端点处的数据延拓得更加准确, 从而使得三次样条曲线在端点处不会发生大的摆动, 实现EMD分解的固有模态函数(IMF)更加准确可靠. 通过对仿真信号的研究表明, 基于PSO-SVM 方法的延拓方法能够很好地抑制了分解的端点效应.     

基于改进的经验模态分解的时间序列匹配算法

针对经验模态分解（EMD）的不足之处,对原有EMD方法中利用上下包络的乎均值得到平均包络进行了改进,采用三次样条对连续极值点的平均值进行插值获得乎均包络。通过这种方式,增加了近似极值点,在“筛”过程的每次循环中,只需要一次而不是两次祥务插值,缓解了“逆冲”和“欠冲”现象,改进了EMD方法,然后引用改进的EMD方法降低序列的维度,并用K均值算法实现模式匹配．实验结果表明,提出的在对EMD进行改进的基础上实现模式匹配的方法,优于传统的基于小波的模式匹配方法。     

基于平均包络线匹配算法的EMD端点效应分析及在股价趋势分解中的应用

经验模式分解(EMD)能够有效获得非平稳非线性信号的时频特征,但传统的EMD分解算法存在严重的端点效应. 在深入研究和分析EMD算法的基础上,提出了一种基于波形匹配的端点效应处理方案,通过计算波形匹配度, 在平均包络线内部寻找与其端部变化趋势最为接近的子波,并用这段子波代替平均包络线的边缘部分, 使处理后的平均包络线极大地接近真实包络线,并把这种端点效应处理方案的EMD分解算法应用到实际的股票市场价格趋势分解中.实验结果表明,与经典的EMD边界延拓算法相比,本文提出的算法能更有效地抑制EMD分解时的边界效应, 分解得到的固有模式函数更能体现模拟信号真实的频率、幅值信息.应用实验表明:与现有方法相比,该方法更能提高预测精度.     

面向入侵检测系统的Deep Belief Nets模型

连续的网络流量会导致海量数据问题,这为入侵检测提出了新的挑战。为此,提出一种面向入侵检测系统的深度信念网络(deep belief nets oriented to the intrusion detection system, DBN-IDS）模型。首先,通过无监督的、贪婪的算法自底向上逐层训练每一个受限玻尔兹曼机(restricted Boltzmann machine,RBM)网络,使得大量高维、非线性的无标签数据映射为最优的低维表示；然后利用带标签数据被附加到顶层,通过反向传播(back-propagation,BP）算法自顶向下有监督地对RBM网络输出的低维表示进行分类,并同时对RBM网络进行微调；最后,利用NSL-KDD数据集对模型参数和性能进行了深入的分析。实验结果表明,DBN-IDS分类效果优于支持向量机(support vector machine,SVM)和神经网络(neural network，NN),适用于高维、非线性的海量入侵数据的分类处理。     

Walsh码软扩频信号降噪算法

针对低信噪比(signal-to-noise ratio, SNR)条件下, Walsh码软扩频信号盲解扩以及多址信号盲分离难以实现的问题,提出一种Walsh码软扩频信号降噪算法。首先,采用经验模态分解(empirical mode decomposition, EMD)算法将Walsh码软扩频信号分解为有限个本征模态函数(intrinsic mode function, IMF),分界点位置可通过Walsh码软扩频信号和噪声的IMF自相关函数收敛速度的差异进行判断。然后,采用小波软阈值滤波算法处理分界点之前的IMF。最后,利用处理后的低阶IMF和分界点后的IMF重构Walsh码软扩频信号,减少由于降噪造成的信号损失。仿真结果表明,在一定低SNR范围内,降噪算法以较低误码率(bit error rate, BER)实现解调,信号损失较少。     

基于“当前”统计模型的高动态GPS信号跟踪算法

提出了一种新的高动态GPS信号跟踪算法,基于＂当前＂统计模型对载波瞬时相位过程、瞬时频率及瞬时频率变化率过程进行建模;在此基础上,结合系统状态方程为线性而量测方程为非线性的特性,采用简化UKF算法对信号参数进行估计,同时采用EKF和标准UKF滤波算法在同等条件下对信号参数进行估计比对。仿真结果表明对信号参数所建模型能真实反映信号的过程,验证了所建模型的正确性的;简化UKF参数估计结果不仅精度与标准UKF相当且具有较高的运算效率,能更好满足高动态下的实时性需求。

Abstract：

A new high dynamic GPS signal tracking algorithm was proposed.Based on current statistical model,a new model for the GPS signal transitional process was established.Since the system involved a linear state equation and a nonlinear measurement equation,a simple Unscented Kalman Filter （UKF） algorithm was applied to estimation the signal parameters.For comparison purpose,an Extended Kalman Filter （EKF） algorithm and a standard UKF algorithm were also applied to the system to illustrate the effectiveness of the proposed filtering algorithm.Simulation results show that the new model can precisely represent the real signal process,and the simple UKF algorithm achieves high precision comparable to standard UKF with higher operation efficiency.Therefore,the new model and the proposed algorithm can satisfy real time requirement in high dynamic environment.     

基于集成智能方法的电力短期负荷预测

将自组织(SOM)和反向传播(BP)两种神经网络结合起来, 并使用模糊理论, 建立了一种基于集成智能方法的日负荷预测智能模型, 该模型首先利用SOM网络的竞争学习能力将历史数据分成若干类别从而找出与预测日同类型的预测类别. 然后, 把温度、日类型等不确定性扰动因素分离出去, 利用BP算法的非线性函数逼近功能, 完成电力负荷的基本分量部分的预测工作. 在处理温度、天气情况、日类型等不确定因素对负荷的影响时, 采用模糊逻辑理论对负荷基本分量进行修正. 提出了一种基于进化树的自组织神经网络算法(SOETA), 该算法是一种无监督基于二叉树的自组织特征映射网络模型, 采用进化思想进行无监督学习, 具有灵活的拓扑结构和精确的模式识别. 本文以2007年厦门市的电力负荷数据为例, 试验结果表明, SOETA+BP+模糊理论的预测精度最优, 有效提高了电力短期负荷预测精度.     

基于EMD-PSR-LSSVM的城市燃气管网短期负荷预测

城市燃气管网短期负荷预测对燃气调度系统的安全与稳定具有重要意义. 为了提高城市燃气管网短期负荷预测精度,建立了基于经验模态分解(EMD)-相空间重构(PSR)-最小二乘支持向量机(LSSVM)的组合预测模型. 首先,运用EMD算法把原始非线性时间序列分解为互不耦合的模态分量,并采用PSR算法确定LSSVM建模中各个分量的输入输出结构; 其次,运用PSO算法对LSSVM建模中的参数进行优化,使用训练好的LSSVM模型对各个IMF分量进行回归预测; 最后运用该组合模型对郑州市燃气管网负荷进行短期预测.结果表明:与LSSVM回归预测和BP神经网络预测模型相比,本文提出的组合模型的预测精度更高,是一种更为有效的城市燃气管网短期负荷预测方法.