High variation subarctic topsoil pollutant concentration prediction using neural network residual kriging

Abstract

The work deals with the application of neural networks residual kriging (NNRK) to the spatial prediction of the abnormally distributed soil pollutant (Cr). It is known that combination of geostatistical interpolation approaches (kriging) and neural networks leads to significantly better prediction accuracy and productivity. Generalized regression neural networks and multilayer perceptrons are classes of neural networks widely used for the continuous function mapping. Each network has its own pros and cons; however both demonstrated fast training and good mapping possibilities. In the work, we examined and compared two combined techniques: generalized regression neural network residual kriging (GRNNRK) and multilayer perceptron residual kriging (MLPRK). The case study is based on the real data sets on surface contamination by chromium at a particular location of the subarctic Novy Urengoy, Russia, obtained during the previously conducted screening. The proposed models have been built, implemented and validated using ArcGIS and MATLAB environments. The networks structures have been chosen during a computer simulation based on the minimization of the RMSE. MLRPK showed the best predictive accuracy comparing to the geostatistical approach (kriging) and even to GRNNRK.

Prediction of nickel concentration in peri-urban and urban soils using hybridized empirical bayesian kriging and support vector machine regression

“…The hybrid model's marginal errors (RMSE and MAE) (EBK SVMR) are two times lower. Similarly, Sergeev et al 34 recorded 0.28 (R 2 ) for the hybrid model developed (multi-layer perceptron residual kriging), compared to 0.637 (R 2 ) for Ni in the current study. The prediction accuracy level of this model (EBK SVMR) is 63.7%, as opposed to 28% obtained by Sergeev et al 34 .…”

“…Recently, there has been a new DSM trend that fosters the combination of geostatistics and MLA in mapping and prediction. Several soil scientists and authors such as Sergeev et al 34 ; Subbotina et al 35 ; Tarasov et al 36 and Tarasov et al 37 have harnessed accurate qualities in geostatistics and machine learning to generate hybrid models that increase the efficiency and quality of the prediction as well as mapping. Some of these hybridizations or combined algorithmic models are artificial neural network-kriging (ANN-RK), multi-layer perceptron residual kriging (MLP-RK), generalized regression neural network residual kriging (GR-NNRK) 36 , artificial neural network-kriging- multilayer perceptron (ANN-K- MLP) 37 and cokriging and gaussian process regression 38 .…”

Prediction of Nickel Concentrations in Urban and Peri-Urban Soils: Application of a Hybridized Empirical Bayesian Kriging and Support Vector Machine Regression Approach

“…The maginal errors (RMSE and MAE) of the hybrid model (EBK SVMR) are two times lower. Similarly, Sergeev et al,23 recorded 0.28 (R 2 ) for the hybrid model developed (multi-layer perceptron residual kriging), compared to 0.637 (R 2 ) for Ni in the current study. The prediction accuracy level of this model (EBK SVMR) is 63.7% as opposed to 28 percent Sergeev et al23…”

“…Recently there have been a new trend in DSM that foster the combination of geostatitics and MLA in mapping and prediction. Several soil scientist and authors such as Sergeev et al, 23 ; Subbotina et al, 24 ; Tarasov et al, 25 and Tarasov et al, 26 have harness accurate qualities in geostatistics and machine learning to generate hybrid models that increase the efficiency and quality of the prediction as well as mapping. Some of these hybrization or combined algrorithmic models are artificial nueral network-kriging (ANN-RK), multi-layer perceptron residual kriging (MLP-RK), generalized regression neural network residual kriging (GR-NNRK) 25 and artificial nueral network-kriging-multilayer perceptron (ANN-K-MLP) .…”

“…Some of these hybrization or combined algrorithmic models are artificial nueral network-kriging (ANN-RK), multi-layer perceptron residual kriging (MLP-RK), generalized regression neural network residual kriging (GR-NNRK) 25 and artificial nueral network-kriging-multilayer perceptron (ANN-K-MLP) . 26 According to Sergeev et al, 23 the act of combining various modelling techniques has the potential to eliminate flaws and increase the efficiency of the hybrid model produced over the single models from which it was developed. Against this backdrop, this new paper deem it necessary to apply a combined algorithm from geostatistic and MLA to develop the best hybridized model to predict the enrichment of Ni in the urban and peri-urban area.…”