Time series prediction techniques have been used in many real-world applications such as financial market prediction, electric utility load forecasting , weather and environmental state prediction, and reliability forecasting. The underlying system models and time series data generating processes are generally complex for these applications and the models for these systems are usually not known a priori. Accurate and unbiased estimation of the time series data produced by these systems cannot always be achieved using well known linear techniques, and thus the estimation process requires more advanced time series prediction algorithms. This paper provides a survey of time series prediction applications using a novel machine learning approach: support vector machines (SVM). The underlying motivation for using SVMs is the ability of this methodology to accurately forecast time series data when the underlying system processes are typically nonlinear, non-stationary and not defined a-priori. SVMs have also been proven to outperform other non-linear techniques including neural-network based non-linear prediction techniques such as multi-layer perceptrons.The ultimate goal is to provide the reader with insight into the applications using SVM for time series prediction, to give a brief tutorial on SVMs for time series prediction, to outline some of the advantages and challenges in using SVMs for time series prediction, and to provide a source for the reader to locate books, technical journals, and other online SVM research resources.

Time Series Prediction Using Support Vector Machines: A Survey

[1] Daily water demand forecasts are an important component of cost-effective and sustainable management and optimization of urban water supply systems. In this study, a method based on coupling discrete wavelet transforms (WA) and artificial neural networks (ANNs) for urban water demand forecasting applications is proposed and tested. Multiple linear regression (MLR), multiple nonlinear regression (MNLR), autoregressive integrated moving average (ARIMA), ANN and WA-ANN models for urban water demand forecasting at lead times of one day for the summer months (May to August) were developed, and their relative performance was compared using the coefficient of determination, root mean square error, relative root mean square error, and efficiency index. The key variables used to develop and validate the models were daily total precipitation, daily maximum temperature, and daily water demand data from 2001 to 2009 in the city of Montreal, Canada. The WA-ANN models were found to provide more accurate urban water demand forecasts than the MLR, MNLR, ARIMA, and ANN models. The results of this study indicate that coupled wavelet-neural network models are a potentially promising new method of urban water demand forecasting that merit further study.

/pdf/comparison-of-multiple-linear-and-nonlinear-regression-3mqykumdrs.pdf

Comparison of multiple linear and nonlinear regression, autoregressive integrated moving average, artificial neural network, and wavelet artificial neural network methods for urban water demand forecasting in Montreal, Canada

Accurate forecasting of electricity load has been one of the most important issues in the electricity industry. Recently, along with power system privatization and deregulation, accurate forecast of electricity load has received increasing attention. Because of the general nonlinear mapping capabilities of forecasting, artificial neural networks have played a crucial role in forecasting electricity load. Support vector machines (SVMs) have been successfully employed to solve nonlinear regression and time series problems. However, SVMs have rarely been applied to forecast electricity load. This investigation elucidates the feasibility of using SVMs to forecast electricity load. Moreover, simulated annealing (SA) algorithms were employed to choose the parameters of a SVM model. Subsequently, examples of electricity load data from Taiwan were used to illustrate the proposed SVMSA (support vector machines with simulated annealing) model. The empirical results reveal that the proposed model outperforms the other two models, namely the autoregressive integrated moving average (ARIMA) model and the general regression neural networks (GRNN) model. Consequently, the SVMSA model provides a promising alternative for forecasting electricity load.

Support Vector Machines with Simulated Annealing Algorithms in Electricity Load Forecasting

Support vector machines with simulated annealing algorithms in electricity load forecasting

Letters: Regularized extreme learning machine for regression problems

Forecasting of air quality parameters is an important topic of atmospheric and environmental research today due to the health impact caused by airborne pollutants existing in urban areas. The support vector machine (SVM), as a novel type of learning machine based on statistical learning theory, can be used for regression and time series prediction and has been reported to perform well with some promising results. The work presented examines the feasibility of applying SVM to predict pollutant concentrations. The functional characteristics of the SVM are also investigated. The experimental comparison between the SVM and the classical radial basis function (RBF) network demonstrates that the SVM is superior to conventional RBF in predicting air quality parameters with different time series.

Air pollutant parameter forecasting using support vector machines

The determination of the proper size of an artificial neural network (ANN) is recognized to be crucial, especially for its practical implementation in important issues such as learning and generalization. In the paper, an effective design method of neural network architectures is presented. The network is firstly trained by a dynamic constructive method until the error is satisfied. The trained network is then pruned by genetic algorithm (GA). The simulation results demonstrate the advantages in generalization and expandability of the proposed method.

Optimal feed-forward neural networks based on the combination of constructing and pruning by genetic algorithms

Modeling of the pollutant concentrations is an important part in the field of atmospheric environment research. Neural network modeling is regarded as a reliable and cost-effective method to achieve such prediction task. In this paper, the principal component analysis technique is used to reduce and orthogonalize input variables of the neural network model, which is established for forecasting the pollutant concentrations in downtown area of Hong Kong. The new approach is demonstrated and validated with two practical cases of predicting the respirable suspended particulate levels in the central area of Hong Kong. The simulation results show that the proposed method is feasible and efficient.

https://www.researchgate.net/profile/L_Wei-Zhen/publication/3959465_Prediction_of_respirable_suspended_particulate_level_in_Hong_Kong_downtown_area_using_principal_component_analysis_and_artificial_neural_networks/links/0046351b6981314de5000000.pdf

A.Y.T. Leung

Papers

Air pollutant parameter forecasting using support vector machines

Optimal feed-forward neural networks based on the combination of constructing and pruning by genetic algorithms

Prediction of respirable suspended particulate level in Hong Kong downtown area using principal component analysis and artificial neural networks