Time-series forecasting through wavelets transformation and a mixture of expert models

Milidiú, Ruy Luiz; Machado, R.J.; Rentería, Raúl P.

doi:10.1016/s0925-2312(98)00120-9

Cited by 42 publications

(23 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For these data sets, the Gaussian kernel performs best among the single SVMs models. The best single SVMs model by using the Gaussian kernel also has better result than the benchmark reported in [13] for the data set C and in [15,19] for the data set D. In the data set A, the best single SVMs model has slightly worse performance than the benchmark reported in [13]. However, among all the methods, the SVMs experts achieve the smallest test error in all the data sets.…”

Section: Santa Fe Competition Time Seriesmentioning

confidence: 82%

“…In Santa Fe data sets A and C, the experimental setup is used as the same as in [13], which is given in Appendix C. There is no validation set, and the parameters of SVMs that produce the smallest NMSE on the test set are used for SVMs, as the same strategy as in [13]. In Santa Fe data set D, the experimental setup is adopted from [15,19].…”

Section: Santa Fe Competition Time Seriesmentioning

confidence: 99%

“…Based on the same idea of using di erent experts for di erent input regions, Milidiu et al [13] generalize the ME architecture into a two-stage architecture to handle the non-stationarity in the data. As shown in Fig.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Support vector machines experts for time series forecasting

2003

View full text Add to dashboard Cite

This paper proposes using the support vector machines (SVMs) experts for time series forecasting. The generalized SVMs experts have a two-stage neural network architecture. In the ÿrst stage, self-organizing feature map (SOM) is used as a clustering algorithm to partition the whole input space into several disjointed regions. A tree-structured architecture is adopted in the partition to avoid the problem of predetermining the number of partitioned regions. Then, in the second stage, multiple SVMs, also called SVM experts, that best ÿt partitioned regions are constructed by ÿnding the most appropriate kernel function and the optimal free parameters of SVMs. The sunspot data, Santa Fe data sets A, C and D, and the two building data sets are evaluated in the experiment. The simulation shows that the SVMs experts achieve signiÿcant improvement in the generalization performance in comparison with the single SVMs models. In addition, the SVMs experts also converge faster and use fewer support vectors.

show abstract

Section: Santa Fe Competition Time Seriesmentioning

confidence: 82%

Section: Santa Fe Competition Time Seriesmentioning

confidence: 99%

See 1 more Smart Citation

Support vector machines experts for time series forecasting

2003

View full text Add to dashboard Cite

show abstract

“…The time-frequency decomposition of wavelet analysis has been successfully applied in image processing [19], signal de-nosing [24] and condition monitoring [20]. Wavelet combined with some expert models such as neural networks was also used efficiently in stochastic time-series forecasting [21]. In the electrical engineering field, the wavelet transform has been successfully used to forecast electrical load and electricity price.…”

Section: Introductionmentioning

confidence: 99%

Wind power prediction interval estimation method using wavelet-transform neuro-fuzzy network

Cai

Zhang

2015

IFS

View full text Add to dashboard Cite

Abstract. Wind power point forecasting is the primary method to deal with its uncertainty. However, in many applications, the probabilistic interval of wind power is more useful than traditional point forecasting. Methods to determine the probabilistic interval of wind power point forecasting value is very essential to power system operations. Based on the bootstrap method, this paper proposed a wavelet transform combined with a neuro-fuzzy network model to estimate the prediction interval of wind power. In the model, to account for the ramp event of wind power series, a wavelet-based ramp event was used and the moving block bootstrap method, which considers the dependence of wind power series, was used to construct sampling datasets. Then, the bootstrapped datasets were estimated by a neuro-fuzzy network inference system. A case study provided a 90% confidence level of prediction intervals, which was constructed to examine the effectiveness of the model.

show abstract

“…In addition, the distribution of the residual term is altered which can be advantageous, especially if the distribution of the original residual term, ε(k), is non-Gaussian [Ljung, 1999]. Several types of transform have been applied in time series forecasting such as Principle Component Analysis (PCA) [Hiden et al, 1999], Independent Component Analysis (ICA) [Roberts et al, 2004], the Fourier Transform (FT) [Schoukens & Pintelon, 1991], the Wavelet Transform (WT) [Yao et al, 2000] and the Wavelet Packet Transform (WPT) [Saito & Coifman, 1997;Roberts et al, 2004;Milidiú et al, 1999;Nason & Sapatinas, 2001] among others. However, the WT and WPT would seem ideal for time series forecasting as unlike PCA, ICA and the FT, some time information is preserved in the transformed variables.…”

Section: Introductionmentioning

confidence: 99%

A Wavelet Transfer Model for Time Series Forecasting

Fay

Ringwood

2007

Int. J. Bifurcation Chaos

View full text Add to dashboard Cite

This paper is concerned with the case of an exogenous system in which a model is required to forecast a periodic output time series using a causal input. A novel approach is developed in which the wavelet packet transform is taken of both the dependent time series and causal input. This results in two sets of basis dictionaries and requires two bases to be chosen. It is proposed that the best bases to choose are those which maximize the mutual information. Input selection is then implemented by eliminating those coefficients of the selected input basis with low mutual information. As an example, a model is constructed to forecast short-term electrical demand.

show abstract

Time-series forecasting through wavelets transformation and a mixture of expert models

Cited by 42 publications

References 5 publications

Support vector machines experts for time series forecasting

Support vector machines experts for time series forecasting

Wind power prediction interval estimation method using wavelet-transform neuro-fuzzy network

A Wavelet Transfer Model for Time Series Forecasting

Contact Info

Product

Resources

About