Wavelet analysis residual kriging vs. neural network residual kriging

Demyanov, Vasily; Soltani, Skander; Kanevski, Mikhaïl; Canu, Stéphane; Maignan, M.; Savelieva, E.; Timonin, Vadim; Писаренко, В. Ф.

doi:10.1007/s004770000056

Cited by 17 publications

(13 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Nakken (1999) might be the first one who used wavelet analysis to characterize the temporal changes of rainfall-runoff and their relationships. In the following, Demyanov et al (2001) used wavelet analysis and geostatistical tools (kriging) for survaying spatial variation of rainfall and the results were compared with the results of the hybrid model of ANN and kriging. Jayawardena et al (2004) also used wavelet decomposition combined with Markov model to simulate daily rainfall of Chao Phvraya watershed in Thailand.…”

Section: Introductionmentioning

confidence: 99%

A new approach to flow simulation using hybrid models

et al. 2017

View full text Add to dashboard Cite

The necessity of flow prediction in rivers, for proper management of water resource, and the need for determining the inflow to the dam reservoir, designing efficient flood warning systems and so forth, have always led water researchers to think about models with highspeed response and low error. In the recent years, the development of Artificial Neural Networks and Wavelet theory and using the combination of models help researchers to estimate the river flow better and better. In this study, daily and monthly scales were used for simulating the flow of Gamasiyab River, Nahavand, Iran. The first simulation was done using two types of ANN and ANFIS models. Then, using wavelet theory and decomposing input signals of the used parameters, sub-signals were obtained and were fed into the ANN and ANFIS to obtain hybrid models of WANN and WANFIS. In this study, in addition to the parameters of precipitation and flow, parameters of temperature and evaporation were used to analyze their effects on the simulation. The results showed that using wavelet transform improved the performance of the models in both monthly and daily scale. However, it had a better effect on the monthly scale and the WANFIS was the best model.

show abstract

Section: Introductionmentioning

confidence: 99%

A new approach to flow simulation using hybrid models

et al. 2017

View full text Add to dashboard Cite

show abstract

“…Combining geological quantification, Wang et al (1999) applied neural networks and geostatistics for modeling the porosity distribution. Demyanov et al (2001) proposed wavelet analysis residual kriging (WARK) and compared the results of validation with the ones obtained from NNRK. Cellura et al (2008) applied neural kriging for the spatial estimation of wind speed.…”

mentioning

confidence: 99%

“…Especially, hybrid models, based on ANNs and geostatistics, have been developed for modeling nonlinear spatial trends and kriging or geostochastic simulation of residuals (Demyanov et al 1998;Wang et al 1999;Demyanov et al 2001;Cellura et al 2008;Chung et al 2012;Padarian et al 2012;Yeh et al 2013). Demyanov et al (1998) applied neural network residual kriging (NNRK) for the spatial interpolation of climate data.…”

mentioning

confidence: 99%

Estimating Spatial Precipitation Using Regression Kriging and Artificial Neural Network Residual Kriging (RKNNRK) Hybrid Approach

Seo

Kim

Singh

2015

Water Resour Manage

View full text Add to dashboard Cite

A hybrid model, combining regression kriging and neural network residual kriging (RKNNRK), is developed for determining spatial precipitation distribution. The RKNNRK model is compared with current spatial interpolation models, including simple kriging (SK), ordinary kriging (OK), universal kriging (UK), regression kriging (RK) and neural network residual kriging (NNRK). Results show that hybrid models, including RK, NNRK and RKNNRK, performed better than SK, OK and UK, based on the coefficient of efficiency (CE), coefficient of determination (r 2 ), index of agreement (d), mean squared relative error (MSRE), mean absolute error (MAE), root-mean-square error (RMSE), and mean squared error (MSE). Of the six spatial interpolation models, the RKNNRK model was the most accurate, and the NNRK model was the second best.

show abstract

“…These models can be built using interpolation techniques such as moving window averaging or some kriging variants, and in the case of high dimensional trend models, they can be constructed by combining lower dimensional trends (McLennan 2007;Leuangthong et al 2008). Also, the use of neural networks to obtain the mean trend has been proposed in Kanevski et al (1996), andDemyanov et al (2001) proposed the use of wavelets and frames for the same purpose. David (1977) and Cressie (1985) discussed the use of relative semivariograms to account for variance trends that are proportional to the local means squared.…”

Section: Introductionmentioning

confidence: 99%

Detecting and quantifying sources of non-stationarity via experimental semivariogram modeling

Cuba

Leuangthong²,

Ortíz

2011

Stoch Environ Res Risk Assess

View full text Add to dashboard Cite

Conventional geostatistics often relies on the assumption of second order stationarity of the random function (RF). Generally, local means and local variances of the random variables (RVs) are assumed to be constant throughout the domain. Large scale differences in the local means and local variances of the RVs are referred to as trends. Two problems of building geostatistical models in presence of mean trends are: (1) inflation of the conditional variances and (2) the spatial continuity is exaggerated. Variance trends on the other hand cause conditional variances to be over-estimated in certain regions of the domain and under-estimated in other areas. In both cases the uncertainty characterized by the geostatistical model is improperly assessed. This paper proposes a new approach to identify the presence and contribution of mean and variance trends in the domain via calculation of the experimental semivariogram. The traditional experimental semivariogram expression is decomposed into three components: (1) the mean trend, (2) the variance trend and (3) the stationary component. Under stationary conditions, both the mean and the variance trend components should be close to zero. This proposed approach is intended to be used in the early stages of data analysis when domains are being defined or to verify the impact of detrending techniques in the conditioning dataset for validating domains. This approach determines the source of a trend, thereby facilitating the choice of a suitable detrending method for effective resource modeling

show abstract

Wavelet analysis residual kriging vs. neural network residual kriging

Cited by 17 publications

References 1 publication

A new approach to flow simulation using hybrid models

A new approach to flow simulation using hybrid models

Estimating Spatial Precipitation Using Regression Kriging and Artificial Neural Network Residual Kriging (RKNNRK) Hybrid Approach

Detecting and quantifying sources of non-stationarity via experimental semivariogram modeling

Contact Info

Product

Resources

About