Time series prediction using evolving radial basis function networks with new encoding scheme

This paper presents a new encoding scheme for training radial basis function (RBF) networks by genetic algorithms (GAs). In general, it is very difficult to select the proper input variables and the exact number of nodes before training an RBF network. In the proposed encoding scheme, both the architecture (numbers and selections of nodes and inputs) and the parameters (centres and widths) of the RBF networks are represented in one chromosome and evolved simultaneously by GAs so that the selection of nodes and inputs can be achieved automatically. The performance and effectiveness of the presented approach are evaluated using two benchmark time series prediction examples and one practical application example, and are then compared with other existing methods. It is shown by the simulation tests that the developed evolving RBF networks are able to predict the time series accurately with the automatically selected nodes and inputs.     

Time series prediction with improved neuro-endocrine model

The paper is focused on improving the performance of neuro-endocrine models with considering the interaction of glands. Comparing to conventional neuro-endocrine models, the concentration of hormone of one gland is modulated by those of others, and the weights of cells are modulated by the improved endocrine system. The interacted equation among all glands is designed and the parameters of them are chosen with theory analysis. Because all the parameters of the model are constants when the system reaches the equilibrium state, particle swarm optimization algorithm is utilized to search the optimal parameters of the model. The theory analysis indicates that the performance of neuro-endocrine model is better than or at least equal to that of corresponding artificial neural network. To indicate the effectiveness of the proposed model, some time series from different research fields, which are used in some literatures, are tested with the proposed model, the results indicate that the proposed model has some good performance.     

Time series prediction using belief network models

We address the problem of generating normative forecasts efficiently from a Bayesian belief network. Forecasts are predictions of future values of domain variables conditioned on current and past values of domain variables. To address the forecasting problem, we have developed a probability forecasting methodology, Dynamic Network Models (DNMs), through a synthesis of belief network models and classical time-series models. The DNM methodology is based on the integration of fundamental methods of Bayesian time-series analysis, with recent additive generalizations of belief network representation and inference techniques.We apply DNMs to the problem of forecasting episodes of apnea, that is, regular intervals of breathing cessation in patients afflicted with sleep apnea. We compare the one-step-ahead forecasts of chest volume, an indicator of apnea, made by autoregressive models, belief networks, and DNMs. We also construct a DNM to analyse the multivariate time series of chest volume, heart rate and oxygen saturation data.     

Bacterially inspired evolving system with an application to time series prediction

This paper explores the synergies between evolutionary computation and synthetic biology, developing an in silico evolutionary system that is inspired by the behavior of bacterial populations living in continuously changing environments. This system creates a 3D environment seeded with a simulated population of bacteria that eat, reproduce, interact with each other and with the environment and eventually die. This provides a 3D framework implementing an evolutionary process. The subject of the evolution is each bacterium's internal process, defining its interactions with the environment. The evolutionary goal is the survival of the population under successive, continuously changing environmental conditions. The key advantage of this bacterial evolutionary system is its decentralized, asynchronous, parallel and self-adapting general-purpose evolutionary process. We describe this system and present the results of an application to the evolution of a bacterial population that learns how to predict the presence or absence of food in the environment by analyzing three input signals from the environment. The resulting populations successfully evolve by continuously improving their fitness under different environmental conditions, demonstrating their adaptability to a fluctuating medium.     

多时间尺度时间序列趋势预测

针对股票、基金等大量时间序列数据的趋势预测问题，提出一种基于新颖特征模型的多时间尺度时间序列趋势预测算法。首先，在原始时间序列中提取带有多时间尺度特征的特征树，其刻画了时间序列，不仅带有序列在各个层次的特征，同时表示了层次之间的关系。然后，利用聚类挖掘特征序列中的隐含状态。最后，应用隐马尔可夫模型（HMM）设计一个多时间尺度趋势预测算法（MTSTPA），同时对不同尺度下的趋势以及趋势的长度作出预测。在真实股票数据集上的实验中，在各个尺度上的预测准确率均在60%以上，与未使用特征树对比，使用特征树的模型预测效率更高，在某一尺度上准确率高出10个百分点以上。同时，与经典自回归滑动平均模型（ARMA）模型和PHMM（Pattern-based HMM）对比，MTSTPA表现更优，验证了其有效性。     

Time series prediction based on data compression methods

We propose efficient (“fast” and low memory consuming) algorithms for universal-coding-based prediction methods for real-valued time series. Previously, for such methods it was only proved that the prediction error is asymptotically minimal, and implementation complexity issues have not been considered at all. The provided experimental results demonstrate high precision of the proposed methods.     

Time series prediction with single multiplicative neuron model

Single neuron models are typical functional replica of the biological neuron that are derived using their individual and group responses in networks. In recent past, a lot of work in this area has produced advanced neuron models for both analog and binary data patterns. Popular among these are the higher-order neurons, fuzzy neurons and other polynomial neurons. In this paper, we propose a new neuron model based on a polynomial architecture. Instead of considering all the higher-order terms, a simple aggregation function is used. The aggregation function is considered as a product of linear functions in different dimensions of the space. The functional mapping capability of the proposed neuron model is demonstrated through some well known time series prediction problems and is compared with the standard multilayer neural network.     

基于多模态信息融合的时间序列预测模型

针对传统单因子模型无法充分利用时间序列相关信息,以及这些模型对时间序列预测准确性和可靠性较差的问题,提出一种基于多模态信息融合的时间序列预测模型——Skip-Fusion对多模态数据中的文本数据和数值数据进行融合。首先利用BERT（Bidirectional Encoder Representations from Transformers）预训练模型和独热编码对不同类别的文本数据进行编码表示;再使用基于全局注意力机制的预训练模型获得多文本特征融合的单一向量表示;然后将得到的单一向量表示与数值数据按时间顺序对齐;最后通过时间卷积网络（TCN）模型实现文本和数值特征的融合,并通过跳跃连接完成多模态数据的浅层和深层特征的再次融合。在股票价格序列的数据集上进行实验,Skip-Fusion模型的均方根误差（RMSE）和日收益（R）分别为0.492和0.930,均优于现有的单模态模型和多模态融合模型的结果,同时在可决系数（R-Squared）上取得了0.955的拟合优度。实验结果表明,Skip-Fusion模型能够有效进行多模态信息融合并具有较高的预测准确性和可靠性。     

基于滑动窗口预测的水文时间序列异常检测

针对水文时间序列分析与决策中存在的数据质量问题,提出了基于滑动窗口预测的水文时间序列异常检测算法。首先基于滑动窗口对时间序列进行子序列分割,再以子序列为基础建立预测模型对未来值进行预测,并将预测值和实测值间差异范围大于预设阈值的序列点判定为异常。探讨了算法中的滑动窗口和参数设置,并以实例数据对算法进行了验证。实验结果表明,所提算法不仅能够有效挖掘出水文时间序列中的异常点,而且将异常检测的灵敏度和特异度分别提高到80%和98%以上。     

Fast evolving neuro-fuzzy model and its application in online classification and time series prediction

A strategy for improving speed of the previously proposed evolving neuro-fuzzy model (ENFM) is presented in this paper to make it more appropriate for online applications. By considering a recursive extension of Gath?CGeva clustering, the ENFM takes advantage of elliptical clusters for defining validity region of its neurons which leads to better modeling with less number of neurons. But this necessitates the computing of reverse and determinant of the covariance matrices which are time consuming in online applications with large number of input variables. In this paper a strategy for recursive estimation of singular value decomposition components of covariance matrices is proposed which converts the burdensome computations to calculating reverse and determinant of a diagonal matrix while keeping the advantages of elliptical clusters. The proposed method is applied to online detection of epileptic seizures in addition to prediction of Mackey?CGlass time series and modeling a time varying heat exchanger. Simulation results show that required time for training and test of fast ENFM is far less than its basic model. Moreover its modeling ability is similar to the ENFM which is superior to other online modeling approaches.     

Time series prediction using Lyapunov exponents in embedding phase space

This paper describes a novel method of chaotic time series prediction, which is based on the fundamental characteristic of chaotic behavior that sensitive dependence upon initial conditions (SDUIC), and Lyapunov exponents (LEs) is a measure of the SDUIC in chaotic systems. Because LEs of chaotic time series data provide a quantitative analysis of system dynamics in different embedding dimension after embedding a chaotic time series in different embedding dimension phase spaces, a novel multi-dimension chaotic time series prediction method using LEs is proposed in this paper. This is done by first reconstructing a phase space using chaotic time series, then using LEs as a quantitative parameter to predict an unknown phase space point, after transferring the phase space point to time domain, the predicted chaotic time series data can be obtained. The computer simulation result of simulation showed that the proposed method is simple, practical and effective.     

基于趋势点状态模型的时间序列预测算法

针对传统的时间序列线性预测算法对时间序列的线性程度要求高,而非线性方法一般建模复杂且计算量大,提出了一种基于趋势点状态模型的时间序列预测算法.该算法无须考虑时间序列是否具有显著线性特征,通过序列间耦合度挖掘时间序列上的相似子序列,找出相对应的相似序列趋势点,建立趋势点状态模型并求出预测值.算法建模简单,复杂度较低.通过模拟实验,结果表明该算法性能良好,尤其对具有周期性的时间序列预测精度很高.     

在线模糊最小二乘支持向量机的时间序列预测

基于模糊最小二乘支持向量机和在线学习算法,提出了一种模糊最小二乘支持向量机的增量式算法。传统最小二乘支持向量机引入模糊加权系数后,有效地提高了其抗噪性能。同时利用递推的核函数计算方法增强了该算法的在线学习能力。仿真结果表明,这一算法在运算精度和运算速度上都优于传统的支持向量机算法。     

LS-SVM时间序列预测 ——免疫文化基因算法进行LS-SVM参数选优

针对最小二乘支持向量机（LS-SVM）在时间序列预测中的参数不确定问题,在训练阶段,使用结合了全局搜索和局部搜索的免疫文化基因算法来进行参数寻优。实验中通过对Lorenz时间序列和建筑能耗的两组预测实验,对比了免疫文化基因算法、遗传算法和网格搜索算法对LS-SVM参数的优化效果,证明了免疫文化基因算法的优化效果最好,且LS-SVM的预测精度比支持向量机（SVM）和BP网络预测都要高。     

基于模体演化与社区一致性的时序链路预测方法

针对动态有向网络中的时序链路预测问题,充分分析动态有向网络中微观结构三元组模体的演化规律,使用指数平滑法季节加法（Holter-Winter-Additive）时序分析方法预测三元组模体的转换概率,引入牛顿法寻求时序分析方法中的最优参数;同时考虑到节点的社区属性对链路预测产生的影响,定义模体内节点的社区结构一致性重要指标,对三元组模体的影响力进行评估。基于此,首先使用时间序列分析方法对模体的转换概率进行预测,进而结合模体社区结构一致性的指标提出一种新的链路预测方法。使用不同的方法在三个真实的有向网络中进行验证,实验结果显示该方法能够达到更好的链路预测效果。     

An evolving recurrent interval type-2 intuitionistic fuzzy neural network for online learning and time series prediction

The prediction of time series has both the theoretical value and practical significance in reality. However, since the high nonlinear and noises in the time series, it is still an open problem to tackle with the uncertainties and fuzziness in the forecasting process. In this article, an evolving recurrent interval type-2 intuitionistic fuzzy neural network (eRIT2IFNN) is proposed for time series prediction and regression problems. The eRIT2IFNN employs interval type-2 intuitionistic fuzzy sets to enhance the modeling of uncertainties by intuitionistic evaluation and noise tolerance of the system. In the eRIT2IFNN, the antecedent part of each fuzzy rule is defined using intuitionistic interval type-2 fuzzy sets, and the consequent realizes the Takagi–Sugeno–Kang type fuzzy inference mechanism. In order to utilize the prior knowledge including intuitionistic information, a local internal feedback is established by feeding the rule firing strength of each rule to itself eRIT2IFNN is fully adaptive to the evolving of sequence data by online learning of structure and parameters. A modified density-based clustering is implemented for the structure learning, where both densities and membership degrees are involved to determine the fuzzy rules. Performance of eRIT2IFNN is evaluated using a set of benchmark problems and compared with existing fuzzy inference systems. Moreover, the eRIT2IFNN is tested for identification of dynamics under both noise-free and noisy environments. Finally, a group of practical financial price-tracking problems including high-frequency data of financial future, commodity future and precious metal are used for the evaluation of the proposed inference system.     

Time series analysis and prediction using gated experts with application to energy demand forecasts

In the analysis and prediction of real world systems two of the key problems are nonstation arity (often in the form of switching between regimes) and overfitting (particularly serious for noisy processes). This article addresses these problems using gated experts consisting of a nonlinear gating network and several also nonlinear competing experts. Each expert learns to predict the conditional mean and each expert adapts its width to match the noise level in its regime. The gating network learns to predict the probability of each expert given the input. This article focuses on the case where the gating network bases its decision on infor mation from the inputs. This can be contrasted to hidden Markov models where the decision is based on the previous state s i e on the output of the gating network at the previous time step as well as to averaging over several predictors. In contrast, gated experts soft partition the input space. This article discusses the underlying statistical assumptions, derives the weight update rules and compares the performance of gated experts to standard methods on three time series: 1 - a computer generated series obtained by randomly switching between two nonlinear processes; 2 - a time series from the Santa Fe Time Series Competition the light intensity of a laser in chaotic state; and 3 - the daily electricity demand of France (a real world multivariate problem with structure on several timescales). The main results are (1) the gating network correctly discovers the different regimes of the process (2) the widths associated with each expert are important for the segmentation task and they can be used to characterize the subprocesses and (3) there is less overfitting compared to single networks homogeneous multilayer perceptrons since the experts learn to match their variances to the local noise levels. This can be viewed as matching the local complexity of the model to the local complexity of the data.     

Efficient polygon decomposition into singular and regular regions via Voronoi diagrams

A new polygon decomposition into regular and singular regions is defined; it is a concept that is useful for skeleton extraction and part analysis of elongated shapes. Polygon regions that are narrow according to the Voronoi diagram of the polygon are extended through the boundary that is adjacent and quasiparallel. Regular regions are the narrow ones surrounded by smooth quasiparallel contour segments, while singular regions are the polygon regions that are not regular. We present an efficient algorithm to calculate the decomposition and make a comparative study with previous algorithms.Received: 21 December 2001, Accepted: 10 January 2003, Published online: 6 June 2003     

基于动态进化图的行人轨迹预测

行人轨迹预测在自动驾驶和社交机器人等领域有着广泛的应用.对行人间复杂的交互关系进行有效建模是提高轨迹预测准确性的关键问题.然而,基于图神经网络的方法建模行人间的复杂交互时,存在行人间交互关系不会随着时间推移而改变,并且图模型无法自适应地调整网络参数,导致预测轨迹与真实轨迹偏差较大.为此,提出基于动态进化图的行人轨迹预测方法,设计动态特征更新(DFU)以定义行人间的动态特性,对行人间动态交互进行建模以构建时间域的网络动态性,提升对行人间复杂交互关系建模的能力.采用进化图卷积单元优化编码器,灵活进化图模型网络参数,增强图模型的自适应能力.研究结果表明,在预测8个时间步长下,与STGAT模型相比,所提出模型在两个公开数据集(ETH和UCY)上取得了更好的性能,平均位移误差降低12.26%,最终位移误差降低14.10%.     

Landfill gas emission prediction using Voronoi diagrams and importance sampling

Municipal solid waste (MSW) landfills are among the nation's largest emitters of methane, a key greenhouse gas, and there is considerable interest in quantifying the surficial methane emissions from landfills. There are limitations in obtaining accurate emissions data by field measurements, and in characterizing an entire landfill with only a few such emissions measurements. This paper proposes an emissions prediction approach using numerous ambient air volatile organic compound (VOC) measurements above the surface of a landfill that are more easily obtained. Many large landfills are already collecting ambient air methane data based on existing regulations. The proposed method is based on the inverse solution of the standard Gaussian dispersion equations. However, only the VOC concentrations and locations are required. The locations of maximum likelihood of the point sources are predicted using Voronoi diagrams, and importance sampling is performed to further refine the locations. Point source strengths are calculated using non-negative least squares, and the point emission rates are then summed to give the total landfill emission rate. The proposed method is successfully demonstrated on a series of four landfill case studies. Three hypothetical landfills were selected for validation studies by forward and backward solution of the dispersion equations. The fourth case study is an active central Florida MSW landfill. The proposed method shows promise in accurately and robustly predicting landfill gas emissions, and requires only measured ambient VOC concentrations and locations.