Principal Component Wavelet Neural Network Research Articles

Quantitative structure-critical micelle concentration modeling of anionic gemini surfactants, comparison of MLR, PLS, WNN, and ANFIS models with eigenvalue and correlation ranking methods

Unique properties of gemini surfactants have attracted the attention of the cosmetics, paints, textile, and detergents industries. Therefore, design of new gemini surfactants is considered important. In this work, the critical micelle concentration of 85 anionic gemini surfactants was predicted by structural descriptors for the first time. Genetic algorithm-multiple linear regression (GA-MLR), genetic algorithm-partial least squares (GA-PLS), principal component-adaptive neuro-fuzzy inference systems (PC-ANFIS), and principal component-wavelet neural network (PC-WNN) were used as linear and nonlinear models. Between linear models, the PLS method had better root mean square errors (RMSE) and mean relative error (MRE) values than MLR model. The eigenvalue and correlation ranking methods were used to select the best PCs in nonlinear models such as PC-ANFIS and PC-WNN. The correlation ranking selection method has better results in both the ANFIS and WNN models. WNN has better results with respect to other models. The RMSE of the training, test, and validation sets for the best PC-WNN model with eight inputs were 0.250, 0.359, and 0.310, respectively.

Comparison of magnetic Fe3O4/chitosan and arginine-modified magnetic Fe3O4/chitosan nanoparticles in simultaneous multidye removal: Experimental design and multicomponent analysis

In this study, multidye biosorption of Titan yellow (TY), Fuchsine acid (FA) and Indigo carmine (IC) in ternary mixture onto magnetic Fe3O4/chitosan nanoparticles (MFe3O4/CS NPs) and arginine-modified MFe3O4/CS NPs was investigated. Simultaneous determination of TY, IC, and FA in ternary solutions was performed by principal component-wavelet neural network (PC-WNN). Characterizations of the synthesized particles were performed by scanning electron microscopy (SEM), elemental analysis (EDAX), thermo-gravimetric analysis (TGA), vibrating sample magnetometer (VSM), Fourier transform infrared (FT-IR) spectroscopy and BET surface area analysis. Result VSM indicate the saturation magnetization values of bare Fe3O4 and arginine-modified magnetic chitosan nanoparticles were about 55.00, 23.04 emu/g, respectively. Also result of BET analysis showed chitosan, MFe3O4/CSNPs, and arginine-modified are nonporous and specific surface area for arginine-modified is more than chitosan, MFe3O4/CSNPs. Response surface methodology (RSM) with Box-Behnken design was used to optimize experimental parameters (biosorbent dose, pH, and time). At optimum conditions, maximum total biosorption capacity of MFe3O4/CSNPs and arginine-modified MFe3O4/CSNPs for anionic dyes was 0.385 and 0.681 mmol g−1, respectively. Arginine-modified magnetic Fe3O4/CSNPs biosorption capacity was twice better than that of magnetic Fe3O4/CSNPs. The biosorption processes were kinetically followed by a pseudo-second order model. Fourteen isotherms were also fitted to biosorption equilibrium data.

Electrical resistivity imaging inversion: An ISFLA trained kernel principal component wavelet neural network approach

The traditional artificial neural network (ANN) inversion of electrical resistivity imaging (ERI) based on gradient descent algorithm is known to be inept for its low computation efficiency and does not ensure global convergence. In order to solve above problems, a kernel principal component wavelet neural network (KPCWNN) trained by an improved shuffled frog leaping algorithm (ISFLA) method is proposed in this study. An additional kernel principal component (KPC) layer is applied to reduce the dimensionality of apparent resistivity data and increase the computational efficiency of wavelet neural network (WNN). Meanwhile, a novel ISFLA algorithm is adopted for improving the learning ability and inversion quality of WNN. In the proposed ISFLA, a hybrid LC mutation attractor is used to enhance the exploitation ability and a differential updating rule is used to enhance the exploration ability. Four groups of experiments are considered to demonstrate the feasibility of the proposed inversion method. The inversion results of the synthetic and field examples show that the introduced method is superior to other algorithms in terms of prediction accuracy and computational efficiency, which contribute to better inversion results.

Simultaneous multidye biosorption by chemically modified Sargassum glaucescens: Doehlert optimization and kinetic, equilibrium, and thermodynamic study in ternary system

ABSTRACTMultidye biosorption of Sunset yellow (SY), Eosin yellow (EY) and Indigo carmine (IC) dyes onto chemically modified biomass of brown alga Sargassum glaucescens was studied. Principal component-wavelet neural network (PC-WNN) was applied for the simultaneous determination of anionic dye concentrations in ternary solutions. Experimental parameters were optimized using response surface methodology (RSM) with a Doehlert design. The optimal biosorption conditions were identified as biosorbent dosage 0.2 g L−1, pH 3, and time 25 min. The maximum total biosorption capacity was 0.102 mmol g−1. The Hill isotherm was the most suitable adsorption models in single and ternary systems. Scanning electron microscopy (SEM) and Fourier transform infrared (FT-IR) analysis confirmed possible interactions between biosorbent surface and dye molecules.

Application of chemometric methods in modeling of competitive multidye biosorption from ternary system

Simultaneous biosorption properties of three cationic dyes (Methylene blue, Crystal violet and Safranin) on Sargassum glaucescens were studied. In the most of previous papers in the field of dye biosorption, one dye or dyes with nearly separate spectra were used and dye concentration was determined by Beer’s law at different λ max. Significant of this study is application of dyes with highly overlapped spectra (as many real situations) that their concentrations can be determined by chemometric methods. Plackett–Burman design was applied to identify the most significant factors, Box–Behnken design was used to determine optimal conditions, and principal component-wavelet neural network was used for the simultaneous determination of dye concentrations in ternary solutions. The optimum biosorption conditions were determined as dye concentration 10−4 mol L−1, biosorbent dosage 0.1 g L−1 and biosorbent particle size 0.188 µm. At this condition, maximum biosorption capacity was 0.80 mmol g−1. The biosorption process was slightly slower in the ternary system comparing with single system which was related to competition phenomena between dyes. It was found that the overall biosorption data were described by the pseudo-second-order kinetic model. Fourteen isotherm models were applied to experimental data, and it was concluded that Hill model had the best correlation.

Prediction Model of the Cement Strength Based on the Principal Component Wavelet Neural Network

In this paper, we study the integration of principal component analysis and wavelet neural net work to establish the principle component wavelet neural network model, and apply the model to predict the strength of cement, and using the average prediction error of the strength of cement by the model to compare with the results through the BP neural network model and the principal component neural network model, the average prediction error in this paper is the smallest. Therefore, the method in this paper has guiding significance to control the quality of the cement production.

Differential evolution trained kernel principal component WNN and kernel binary quantile regression: Application to banking

In this paper, two novel kernel based soft computing techniques viz., Differential Evolution trained Kernel Principal Component Wavelet Neural Network (DE-KPCWNN) and DE trained Kernel Binary Quantile Regression (DE-KBQR) are proposed for classification. While, the former can solve multi-class classification problems, the latter can solve binary classification problems only. In the proposed DE-KPCWNN technique, Kernel Principal Component Analysis (KPCA) is applied to input data to get Kernel Principal Components, on which we will employ WNN. Then, DE is used to train the resulting KPCWNN. In DE-KBQR we applied Kernel technique on the input data to get Kernel Matrix, on which we will employ BQR. Then, DE is used to train the resulting KBQR. Several experiments are conducted on four bankruptcy datasets, three benchmark datasets and two Credit scoring datasets to assess the effectiveness of the proposed classification techniques. The results indicate that the proposed Soft Computing hybrids for classification are efficient than the existing classification techniques. Out of the two, DE-KBQR performed relatively better compared to DE-KPCWNN on a majority of binary classification problems. This is the significant outcome of this study.

Simultaneous kinetic determination of thiocyanate and sulfide using eigenvalue ranking and correlation ranking in principal component-wavelet neural network

In this work, two toxic compound, sulfide and thiocyanate were determined simultaneously using kinetic spectrophotometry. These anions have shown the catalytic effects on the reaction between iodine and azide. Since the system was nonlinear, a nonlinear model, principal component-wavelet neural network (PC-WNN) was used as the multivariate calibration method. The principal component analysis was used to decrease the dimension of the original matrix. In other words, the scores of the PCs, 5, instead of the original variables, 301, were used as the input for the model. Two methods were used to select the most relevant principal components: eigenvalue ranking and correlation ranking. In this work, eigenvalue and correlation ranking methods have shown better results for thiocyanate and sulfide, respectively, and it can be concluded that these methods are complementary. The WNN has several advantages relative to other types of neural network such as better convergence ability. The data set was divided to calibration, prediction and validation sets. Each set was selected so that the concentrations of the analytes were approximately covered the entire ranges of the analytes. Mean relative error for thiocyanate and sulfide in validation set were 8.5 and 10.6, respectively. Thiocyanate and sulfide can be determined in the range of 60–700 ng ml −1 and 20–400 ng ml −1, respectively. The proposed method was applied for the determination of sulfide and thiocyanate in real samples such as tap, waste and river waters with satisfactory results.

Principal Component‐Wavelet Neural Network as a Multivariate Calibration Method for Simultaneous Determination of Iron, Nickel, and Cobalt

Principal component‐artificial neural network (PC‐ANN) and principal component‐wavelet neural network (PC‐WNN) are applied for simultaneous determination of iron(II), nickel(II), and cobalt(II). A simple and selective spectrophotometric method for simultaneous determination of iron(II), nickel(II), and cobalt(II) based on formation of their complexes with 1‐(2‐pyridylazo)‐2‐naphtol (PAN) in micellar media is described. Although the complexes of Fe(II), Ni(II), and Co(II) with reagent show a spectral overlap, they have been simultaneously determined by PC‐ANN and PC‐WNN. The results obtained by the two methods were compared and it was shown that in PC‐WNN, the convergence speed was faster and the root mean square error of prediction set was also smaller than PC‐ANN. Interference effects of common anions and cations were studied and the proposed method was also applied satisfactorily to the determination of Fe(II), Ni(II) and Co(II) in synthetic samples.

Principal Component‐Wavelet Neural Networks as a New Multivariate Calibration Method

A principal component‐wavelet neural network (PC‐WNN) was proposed as a multivariate calibration method for simultaneous determination of test samples of copper, iron, and aluminum. Principal component analysis has been applied to the data set for dimensionality reduction of the data matrix, and neural network with wavelet function has been employed as the function‐learning method. The data sets consisted of 27 standard solutions, which were randomly divided into training and prediction sets. The WNN architecture and its parameters are optimized based on the minimum value for the root mean square errors of the prediction set to prevent over‐fitting the model. The performance of the constructed model is evaluated by prediction of the metal concentrations in a validation set and also in the alloy samples.