Application of Bayesian Regularized BP Neural Network Model for Trend Analysis, Acidity and Chemical Composition of Precipitation in North Carolina

Abstract

Bayesian regularized back-propagation neural network (BRBPNN) was developed for trend analysis, acidity and chemical composition of precipitation in North Carolina using precipitation chemistry data in NADP. This study included two BRBPNN application problems: (i) the relationship between precipitation acidity (pH) and other ions (NH + 4 , NO − 3 , SO 2− 4 , Ca 2+ , Mg 2+ , K + , Cl − and Na + ) was performed by BRBPNN and the achieved optimal network structure was 8-15-1. Then the relative importance index, obtained through the sum of square weights between each input neuron and the hidden layer of BRBPNN(8-15-1), indicated that the ions' contribution to the acidity declined in the order of NH + 4 > SO 2− 4 > NO − 3 ; and (ii) investigations were also carried out using BRBPNN with respect to temporal variation of monthly mean NH + 4 , SO 2− 4 and NO 3 − concentrations and their optimal architectures for the 1990-2003 data were 4-6-1, 4-6-1 and 4-4-1, respectively. All the estimated results of the optimal BRBPNNs showed that the relationship between the acidity and other ions or that between NH + 4 , SO 2− 4 , NO − 3 concentrations with regard to precipitation amount and time variable was obviously nonlinear, since in contrast to multiple linear regression (MLR), BRBPNN was clearly better with less error in prediction and of higher correlation coefficients. Meanwhile, results also exhibited that BRBPNN was of automated regularization parameter selection capability and may ensure the excellent fitting and robustness. Thus, this study laid the foundation for the application of BRBPNN in the analysis of acid precipitation.

Predictive Abilities of Bayesian Regularization and Levenberg–Marquardt Algorithms in Artificial Neural Networks: A Comparative Empirical Study on Social Data

“…That is, in contrast to conventional network training where an optimal set of weights is chosen by minimizing an error function, the Bayesian approach involves a probability distribution of network weights. As a result, the predictions of the network are also a probability distribution [19,20].…”

Artificial Neural Network Approach for Predicting Cation Exchange Capacity in Soil Based on Physico-Chemical Properties

“…Facing with the dynamic, nonlinear, and uncertain characteristics of soil properties in the complex ecosystem, a recent approach to model PTFs is to use an artificial neural network (ANN). Most researchers have found that ANN performs better than multiple linear regressions (MLRs), with no specific function of relationship to be assumed beforehand, capable of predicting large quantities of complex soil properties precisely that encompass most soils in a survey area (Minasny and McBratney, 2002;Minasny et al, 2004;Tang et al, 2006;Xu et al, 2006). Radial basis function neural network (RBFN) is an important model of feedforward neural networks with a single hidden layer and a linear output layer (Panagou et al, 2007).…”

Trade-Offs in Multi-Purpose Land Use under Land Degradation

“…AHP is appropriate for regional management and planning as it can accommodate conflicting, multi-dimensional, incommensurable, and incomparable set of objectives [131]. Taking sustainability into account, this technique involves the paired comparisons of socio-economic objectives that are considered to be as important as eco-political aspects [133,134]. While AHP is an important member of a general family of multi-criteria decision making, which helps to combine the information from various criteria into a single index of evaluation, RS and GIS can capture a wide range of criteria data that are derived from different multi-spatial, multi-temporal, and multi-scale sources for a time-efficient and cost-effective analysis.…”