Levenberg-Marquardt Back Propagation Neural Network Research Articles

Factors influencing indoor air pollution in buildings using PCA-LMBP neural network: A case study of a university campus

This study investigated indoor air quality (IAQ) and its association with building characteristics and environmental factors in eleven different campus buildings in Gainesville, Florida. Integrated indoor and outdoor sensor systems are built and installed to measure the levels of airborne particulate matter (PM2.5 and PM10), nitrogen dioxide (NO2), ozone (O3), relative humidity (RH), and temperature continuously in 10 min intervals for two weeks for each case building. Twenty building-related characteristics were collected through a walkthrough-based survey, HVAC historical data, and construction drawings. Through these data, a PCA-assisted Levenberg-Marquardt Backpropagation (LMBP) neural network model was developed for rapidly and accurately analyzing and predicting the interaction between IAQ and its affecting factors. Factors with significant contributions to indoor exposure were may mostly be determined by outdoor sources. This is evidenced by the strong associations that were found between indoor PM(2.5-10) and O3 values and their corresponding outdoor values and factors, including distance from the major traffic (DFMT), cracks occurring, outdoor temperature, and humidity. Indoor NO2 concentrations were affected by DFMT, indoor O3, indoor RH, number of air grilles, room volume, and window-to-wall ratio. Also, the comparison shows that the PCA-LMBP model outperforms the traditional BP-ANN and multi-linear regression methods. The average values of 1.34, 2.53, and 4.86 were obtained for the root-mean-square error (RMSE) of PCA-LMBP, BP-ANN, and MLR models, respectively. These results can be accordingly referred for the follow-up studies that analyze IAQ in similar building and environmental conditions.

Brain tumour detection in magnetic resonance imaging using Levenberg–Marquardt backpropagation neural network

Magnetic resonance imaging (MRI) is a high-quality medical image that is used to detect brain tumours in a complex and time-consuming manner. In this study, a back propagation neural network (BPNN) along with the Levenberg–Marquardt algorithm (LMA) is proposed to classify MRIs and diagnose brain tumours in a simple and fast process. The BPNN has 10 neurons in the hidden layer, and the default function of the feedforward feeds is mean squared error (MSE). The LMA is optimized as a multivariable adaptive approach and considerably decreases the MSE of the BPNN, so the errors of the tumour classification are diminished. The proposed method follows four steps including preprocessing, skull removal, feature extraction, and classification. The input MRIs are converted to greyscale, resized, and thresholding is performed in the preprocessing step and followed by skull removal. Morphological operations of closing, opening, and dilation are used to segment abnormal areas in the MRIs, and the opening operator recognizes the tumour more accurately. Using statistical analysis and a grey-level co-occurrence matrix (GLCM) 12 features are extracted from the MRIs and used as the inputs of the BPNN. To evaluate the proposed method, 670 normal and 670 abnormal brain MRIs are used as input data, and the classification is performed in 0.494 s. The accuracy, sensitivity, specificity, precision, dice, recall, and MSE are 98.7%, 97.61%, 99.7%, 97.61%, 98.6%, 97.61%, and 0.005, respectively. The approach is accurate and fast for medical images classification.

MHD williamson nanofluid flow in the rheology of thermal radiation, joule heating, and chemical reaction using the Levenberg–Marquardt neural network algorithm

Various studies have been conducted on the topic of predicting the thermal conductivity of nanofluids. Here, the thermal conductivity of nanofluids is determined using artificial neural networks since this approach is rapid and accurate, as well as cost-effective. To forecast the thermal conductivity of magnetohydrodynamic Williamson nanofluids flow through a vertical sheet, a feed-forward neural network with various numbers of neurons has been evaluated, and the best network based on the performance is selected. The fluid model incorporates the effects of Joule heating, heat generation absorption, thermal radiation, and a chemical reaction (MHD-WNF-HGA). A combination of heat radiation and reactive species improves the energy and solute profiles. The magnetic Reynolds number is assumed to be so small; therefore, the generated magnetic field has no effect. A postulate of similarity variables is used to convert the physical model in the form of nonlinear partial differential equations to an ordinary differential equation system. A supervised Levenberg–Marquardt backpropagation algorithm possesses a multilayer perceptron that is used for training the network, which is one of the top algorithms in machine learning. The bvp4c numerical technique is adopted to build the datasets for the construction of continuous neural network mapping. Flow, energy, and concentration profiles of the fluidic flow are constructed by adjusting several physical quantities such as the Williamson parameter, thermal radiation parameter, magnetic parameter, Eckert number, Darcy number, Brownian motion, and thermophoresis parameter. Analytical techniques such as error histogram graphs and regression-based statistical graphs are used to examine the accuracy of a suggested method. It has been found that the Levenberg–Marquardt backpropagation neural network mappings’ derivation, convergence, authentication, and consistency have been proven. Furthermore, thermal radiation assists the energy distribution to increase smoothly. Fluid velocity drops with the Williamson parameter, whereas thermophoresis impact enhances the strength of the nanofluid density.

BAMCloud: a cloud based Mobile biometric authentication framework

There has been an exponential increase in the number of users switching to mobile banking. Therefore, various countries are adopting biometric solutions as security measures. Biometric technologies provide the potential security framework to make banking more convenient and secure than it has ever been. These technologies are gaining much popularity because of the ease in capturing biometric data in real-time using one’s mobile phone. At the same time, the exponential growth of enrollment in the biometric system produces a massive amount of high-dimensional data. To overcome performance-related issues arising due to the resulting data deluge, this paper aims to propose a distributed mobile biometric system based on a high-performance cluster Cloud. In this paper, a Cloud-based mobile biometric authentication framework (BAMCloud) is proposed that uses dynamic signatures for authentication. The process flow of the BAMCloud system involves capturing data using any handheld mobile device, followed by its storage, preprocessing, and training of the system in a distributed manner over the Cloud. MapReduce has been implemented on the Hadoop platform to reduce the processing time. For model training, The Levenberg-Marquardt backpropagation neural network has been used. It achieves a speed of 8.5 times the original speed and performance of 96.23%. Furthermore, the cost-benefit analysis of the implemented system shows that the cost of implementation and execution of the system is less than the existing ones. The experiments demonstrate that better performance is achieved by implementing the proposed framework as compared to other methods used in recent literature.

MHD Casson Nanofluid in Darcy-Forchheimer Porous Medium in the Presence of Heat Source and Arrhenious Activation Energy: Applications of Neural Networks

ABSTRACT The present study investigates the significance of activation energy during chemical reactions, thermal radiations and temperature gradient of 3-D steady Magnetohydrodynamic Casson nanofluid flow (MHDCNFM) in Darcy−Forchheimer medium over an oscillating disk by using an intelligent numerical-based computing solver through the Levenberg-Marquardt backpropagation neural network scheme (LMBNNS). The adjusted Buongiorno model is utilized to develop the system of partial differential equations (PDEs) for MHDCNFM, and further, by invoking Von Karman similarity transformation, the system of governing PDEs is turned out into ordinary differential equations (ODEs). First, a data set for the magnetohydrodynamic Casson nanofluid flow model (MHDNNFM) is generated for a range of sundry parameters for the radial velocity, tangential velocity, heat distribution and concentration profile by the variations of Casson parameter, magnetic parameter, Brownian motion parameter, thermophoresis parameter, Forchheimer parameter, activation energy parameter, chemical reaction parameter, stretching parameter, porosity parameter and Schmidt number through the Lobatto IIIA technique, and after this, by applying an intelligent computing algorithm through nftool training, testing and validation steps are taken into account to find out the approximate solution for various cases. The designed solver LMBNNS is used to solve the MHDCNFM through mean squared error (MSE), regression, gradient analysis and histogram studies.

Buildings’ internal heat gains prediction using artificial intelligence methods

A large part of energy consumption in homes, offices and commercial spaces is related to Heating, Ventilation and Air-conditioning (HVAC) devices. The effective parameter on the consumption of HVAC systems is internal heat gains that arise from occupants, electric equipment and lighting. In order to reduce the energy consumption of these systems, internal heat gains should be predicted accurately. Since there are few investigations performed on the prediction of internal heat gains, in this paper, three predictive models, namely multiple regression model, Levenberg–Marquardt back-propagation (LM-BP) model and similar days method based on combined weights, have been deployed. By assessing the influential factors on internal heat gains, fundamental theories, structures, equations and parameters of these models are thoroughly proposed. To examine the prediction techniques, an office building in China was considered. It was found that all the proposed models have high accuracy; however, the LM-BP neural network showed the most precision among other models with RMSE = 15.59, MAE = 10.16 and MAPE = 6.35. This model had a higher agreement with the actual internal heat gains compared to the predetermined working programs in the ASHRAE standard 90.1. The proposed models used in this study can lead to providing a theoretical base for scholars and engineers to improve the predictive control of HVAC systems, which plays an important role in enhancing thermal comfort, saving energy of residential buildings.

Classification Framework for Clinical Datasets Using Synergistic Firefly Optimization

Classification is a data mining task, which plays a vital role in clinical diagnosis. Irrelevant features will reduce the classifier accuracy. Only a subset of features in a clinical dataset plays a key role in diagnosing the disease. Thus, selecting the relevant features and training the classifier will improve the classifier accuracy. Clinical datasets are subjected to pre-processing, followed by feature selection and classification. In this work, a framework that uses Synergistic firefly algorithm for feature selection and an ensemble classifier for classification has been designed and implemented. The missing values in the clinical datasets are handled using the k-Nearest Neighbour (k-NN) technique. Min–max normalization is used to normalize the data and the normalized data are split into training and testing sets using tenfold cross-validation. Feature selection has been carried out using a wrapper approach that uses the Synergistic firefly algorithm for generating feature subsets and the Levenberg-Marquardt Back Propagation Neural Network to evaluate the performance of the subsets. The ensemble classifier, consisting of the Levenberg-Marquardt Back Propagation Neural Network, Extreme Learning Machine and Naïve Bayes classifier, is trained using the optimal features selected by the Synergistic firefly algorithm. Experimentation has been carried out using eight clinical datasets from the University of California Irvine (UCI) machine learning repository and it has been inferred that selecting the relevant features improves the classifier accuracy. The results obtained prove the efficiency of the decision-making system developed using Synergistic Firefly and the decision-making system can assist the clinician in decision-making to diagnose diseases.

Levenberg-marquardt backpropagation neural network with techebycheve moments for face detection

Face detection is an intelligent approach used in a variety of applications that identifies human faces in digital images. This work presents a new method which composes of a neural network and Techebycheve transforms for face detection. For feature extraction, Tchebychev transform was applied, in which a discrete Tchebychev transform is given for different sampling patterns and several samples here were performed on color images. A Levenberg-Marquardt backpropagation neural network was applied to the transformed image to find faces in the face detection dataset and FDDB benchmarked database. Model performance was measured based on its accuracy and the best result from the newly proposed method was 98.9%. Simulation results showed that the proposed method handles face detection efficiently.

Water Quality Analysis and Evaluation of Eutrophication in a Swamp Wetland in the Permafrost Region of the Lesser Khingan Mountains, China.

Wuyiling Nature Reserve is located at the edge of a frozen soil area and has abundant vegetation resources. It is an important area for evaluating the impact of frozen soil degradation on the environment. Analyzing the water quality and eutrophication characteristics of different swamps and water bodies can provide a basis for protecting the water environment of frozen soil areas. The pollution characteristics of different swamps and water bodies were analyzed and the Levenberg-Marquardt back-propagation neural network was used to evaluate water quality and eutrophication. Finally, the eutrophication evaluation was compared with the nutritional status index. The results demonstrated that (1) the highest concentrations of the total phosphorus (TP), total nitrogen, ammonia nitrogen (AN), permanganate index (PI), chemical oxygen demand, and chlorophyll were all present in the Tangwang River, which could be caused by the effects of human and agricultural activities along the river. The maximum average concentrations of TP, AN, and PI in the four wetland swamps appeared in the shrub swamp; (2) The water quality category of shrub, grass, and forest swamps in Wuyiling Nature Reserve was Class II, and the water quality category of floating swamps and the Tangwang River was Class I. The nutrient levels of the four marsh swamp wetlands and the Tangwang River were in the mid-range for many nutrients; and (3) the LM-BP neural network model, which was used to evaluate water eutrophication, and the results of a comprehensive evaluation of a nutritional status index were similar, further demonstrating the credibility of the evaluation results.

Pesticide Residue Measurements in Raw Milk Based on Optical Spectrum and Neural Networks

Pesticide residue has become as one of the major food adulterants as a result , rapid, simple, spectrophotometric method for pesticide residue measurement (bifenthrin, glyphosate) in milk sample based on UV-Vis spectrophotometer and Artificial neural network had been developed .The developed system detecting the milk contaminants like pesticide residue and classified it based on absorption spectrum. The hybrid system consists of UV-Vis spectrophotometer and neural package. The detection and classification processes depend on the readout of the UV-Vis spectroscopy to achieve the detection and recognition of all the samples by using Levenberg-Marquardt back propagation (LM-BP) neural network algorithm. The reported classification rate accuracy reached above 99.98%.

Multi‐objective Deng's grey incidence analysis, orthogonal optimization, and artificial neural network modelling in hot‐maceration‐assisted extraction of African cucumber leaves (Momordica balsamina)

AbstractDue to the inherent multiple response characteristics in many biological and separation processes, parameter optimization and modelling is usually a daunting task. The integration of Deng's grey incidence model (GRA) and Taguchi optimization (TM) therefore helps in transforming multiple quality characteristics into a single response presented as the grey relational grade (GRG). This was applied to optimize the multiple quality response characteristics in the maceration‐assisted extraction of African cucumber leaves. Two responses and five design factors were selected with L16(25) layout using signal‐to‐noise ratio as a point prediction feature. Under the optimized conditions, the optimum total phenolic content and antioxidant capacity of 0.8569 mg/ml gallic acid equivalence and 0.9259 mg/ml were achieved, respectively. The mass ratio was the highest contributor (38.2%), whereas the maceration time presented the least contribution (9.8%) to the cumulative response grade (GRG). In the neural network analysis, three models were deployed: Levenberg Marquardt backpropagation neural network (LMNN), gradient descent with adaptive learning rate neural network (GDALRNN), and the resilient back‐propagation neural network (RPNN). A better prediction of hold‐out data was achieved with the GDALRNN model, generating lesser absolute deviation error (MADGDALRNN = 0.099), root mean square error (RMSEGDALRNN = 0.1033), relative mean bias error (rMBEGDALRNN = − 0.24), and highest computational time (CTGDALRNN = 8.8), which is expected of an effective model. Based on the GRG and the signal‐to‐noise ratio, the optimum conditions and the neural network model succinctly provided a benchmark for future assessment of complex relationship among extraction variables, which could form the basis for a potential future scale‐up applications.

Hybrid neurofuzzy investigation of short-term variability of wind resource in site suitability analysis: a case study in South Africa

Energy generation from wind resources is now a mature technology with the ability to compete with traditional energy sources at utility scales in many countries, through the identification of suitable sites. However, beyond site suitability, predicting the wind resource variability of the potentially viable site presents overarching benefits in strategic and operational planning prior to site development. This study, therefore, combines geographical information systems multicriteria decision-making (GIS-MCDM) and hybrid neurofuzzy modeling tools for site suitability and resource variability forecast, respectively, in the Eastern Cape Province of South Africa. The GIS model uses two factors (climatological and environmental), and analytical hierarchical process was used for evaluating criteria degree of influence. Wind resource variability using diurnal satellite-based data for the candidate site was used on the models. Adaptive neurofuzzy inference system models hybrid with genetic algorithm (GA-ANFIS) and particle swarm optimization (PSO-ANFIS) were compared with standalone ANFIS and Levenberg–Marquardt backpropagation neural network (LMBP-ANN) using six statistical measures of error, accuracy, and variability. The GA-ANFIS and PSO-ANFIS accurately model the resource with PSO-ANFIS having lesser computational time compared to GA-ANFIS. However, LMBP-ANN is most robust and resistant in modeling the resource variability among the four models. Hence, wind resource variability investigation on a potentially viable site obtained from the GIS-MCDM model can complement on-site investigations prior to site development. Also, tuning ANFIS with evolutionary algorithms offers improved accuracy over standalone ANFIS model for wind resource forecast and further its robustness in predicting variability of the resource. From our findings, cross-boundary wind resource exploration between South Africa and Lesotho could foster regional interconnectivity.

Computational intelligence of Levenberg-Marquardt backpropagation neural networks to study the dynamics of expanding/contracting cylinder for Cross magneto-nanofluid flow model

In the present investigation, design of integrated numerical computing through Levenberg-Marquardt backpropagation neural network (LMBNN) is presented to examine the fluid mechanics problems governing the dynamics of expanding and contracting cylinder for Cross magneto-nanofluid flow (ECCCMNF) model in the presence of time dependent non-uniform magnetic force and permeability of the cylinder. The original system model ECCCMNF in terms of PDEs is converted to nonlinear ODEs by introducing the similarity transformations. Reference dataset of the designed LMBNN methodology is formulated with Adam numerical technique for scenarios of ECCCMNF by variation of thermophoresis temperature ratio parameter, Brownian motion, suction parameters as well as Schmidt, Prandtl, local Weissenberg and Biot numbers. To calculate the approximate solution for ECCCMNF for different scenarios, the training, testing, and validation processes are conducted in parallel to adapt neural network by reducing the mean square error (MSE) function through Levenberg-Marquardt backpropagation. The comparative studies and performance analyses based on outcomes of MSE, error histograms, correlation and regression demonstrate the effectiveness of designed LMBNN technique.

Improved Levenberg–Marquardt backpropagation neural network by particle swarm and whale optimization algorithms to predict the deflection of RC beams

The aim of this study is to develop a novel computer-aided method for the prediction of the deflection of reinforced concrete beams (DRCB) under concentrated loads. To this end, in the present work, a Levenberg–Marquardt-based backpropagation novel neural network model, optimized by the whale optimization algorithm (WOA), called WOA-LMBPNN, has been developed. Specifically, a neural network, using the Levenberg–Marquardt backpropagation training algorithm with multiple hidden layers, was optimized by the WOA, aiming to obtain higher accuracy in predicting DRCB. For the training of the models, 120 experiments with the geometrical and mechanical properties of concrete beams were compiled using were used as the input parameters. Seven datasets with different number of input variables were investigated to evaluate the effect of the input variables on DRCB. For comparison purposes, another swarm optimization algorithm (i.e., particle swarm optimization—PSO) was also used to optimize the LMBPNN model (i.e., PSO-LMBPNN model). The results obtained by the PSO-LMBPNN and WOA-LMBPNN models are then compared based on the different datasets. Finally, the results revealed the effective role of the WOA, as well as the efficiency and robustness of the new hybrid WOA-LMBPNN model in predicting DRCB.

The use of artificial neural networks for modelling rumen fill

Artificial neural network (ANN) and random forest models for predicting rumen fill of cattle and sheep were developed. Data on rumen fill were collected from studies that reported body weights, measured rumen fill, and stated diets fed to animals. Animal and feed factors that affected rumen fill were identified from each study and used to create a dataset. These factors were used as input variables for predicting the weight of rumen fill. For ANN modelling, a three-layer Levenberg–Marquardt back-propagation neural network was adopted and achieved 96% accuracy in prediction of the weight of rumen fill. The precision of the ANN model’s prediction of rumen fill was higher for cattle (80%) than sheep (56%). On validation, the ANN model achieved 95% accuracy in prediction of the weight of rumen fill. A random forest model was trained using a binary tree-based machine-learning algorithm and achieved 87% accuracy in prediction of rumen fill. The random forest model achieved 16% (cattle) and 57% (sheep) accuracy in validation of the prediction of rumen fill. In conclusion, the ANN model gave better predictions of rumen fill compared with the random forest model and should be used in predicting rumen fill of cattle and sheep.

Prediction of human odour assessments based on hedonic tone method using instrument measurements and multi-sensor data fusion integrated neural networks

A Cyranose 320 (eNose) and a Fast Gas Chromatograph (CG) analyser (zNose™) were used to measure the headspace odour of solid samples from dairy operations. The measurements of both sensors were trained by Levenberg–Marquardt Back-propagation Neural Network (LMBNN) to match human assessments. A trained human panel was used to assess the odours based on hedonic tone method and provide the model targets. A multi-sensor data fusion approach was developed and applied to integrate the eNose and zNose readings for higher predictive accuracy compared to each sensor alone. Principle Component Analysis, Forward Selection, and Gamma Test were applied to reduce the model input dimensions. Measurement fusion models and information fusion model approaches were applied. The information fusion prediction models were shown to be more accurate than all other models, including single instrument models. The information fusion model based on eNose with Gamma Test data reduction + zNose showed the best results of all cases in validation mean square error (0.34 odour units), R value (0.92), probability of the prediction falling within 10% of the target (96%), and probability of the prediction falling within 5% of the target (63%).

A quality prediction model of plastic gears based on an LMBP neural network and the PSO algorithm

ABSTRACT This paper describes a new method for finding optimal process parameter values of plastic gears based on a small experimental dataset. Taking a polyamide 66 gear as an example, the simulation study and the real experiment have been designed based on the Taguchi approach. Then, the Levenberg-Marquardt back propagation (LMBP) neural network is built based on the data generated by the real experiment. This network is also used as the fitness function of the particle swarm optimization algorithm. The results show that, by using optimal parameter values, the comprehensive score of the polyamide 66 gear increases by 8.71%.

Complexity Analysis of a Four-Dimensional Energy-Economy-Environment Dynamic System

A novel four-dimensional energy, economic, and environmental (3E) under energy reduction constraints chaotic system is proposed. The acquisition of environmental quality data is the key to this paper. During the course of the study, we used Bayesian estimation algorithm to calibrate the environmental quality. Based on the official data, the Levenberg–Marquardt backpropagation neural network method was optimized by genetic algorithm to effectively identify the parameters in the 3E system. The research results show that although increasing energy reduction inputs can improve environmental quality, the effect on energy intensity and overall stability of the system is not obvious. When polluting input in the ecological environment system affects its maximum capacity, the environmental system collapses (i.e., the ecological system can no longer purify the environment through the self-circulation process and will eventually die out). Therefore, it is necessary to correctly grasp the ecological environment protection and the relationship between economic developments and explore synergies to promote ecological priorities and green development new ideas.

Compressive Strength Prediction of PVA Fiber-Reinforced Cementitious Composites Containing Nano-SiO2 Using BP Neural Network.

In this study, a method to optimize the mixing proportion of polyvinyl alcohol (PVA) fiber-reinforced cementitious composites and improve its compressive strength based on the Levenberg-Marquardt backpropagation (BP) neural network algorithm and genetic algorithm is proposed by adopting a three-layer neural network (TLNN) as a model and the genetic algorithm as an optimization tool. A TLNN was established to implement the complicated nonlinear relationship between the input (factors affecting the compressive strength of cementitious composite) and output (compressive strength). An orthogonal experiment was conducted to optimize the parameters of the BP neural network. Subsequently, the optimal BP neural network model was obtained. The genetic algorithm was used to obtain the optimum mix proportion of the cementitious composite. The optimization results were predicted by the trained neural network and verified. Mathematical calculations indicated that the BP neural network can precisely and practically demonstrate the nonlinear relationship between the cementitious composite and its mixture proportion and predict the compressive strength. The optimal mixing proportion of the PVA fiber-reinforced cementitious composites containing nano-SiO2 was obtained. The results indicate that the method used in this study can effectively predict and optimize the compressive strength of PVA fiber-reinforced cementitious composites containing nano-SiO2.

A Levenberg–Marquardt Backpropagation Neural Network for the Numerical Treatment of Squeezing Flow With Heat Transfer Model

In this paper, the computational strength in terms of soft computing neural networks backpropagated with the efficacy of Levenberg-Marquard training (NN-BLMT) is presented to study the squeezing flow with the heat transfer model (SF-HTM). The governing system of PDEs is reduced to an equivalent system of nonlinear ODEs using similarity transformations. NN-BLMT dataset for all problem scenarios progresses through the standard Adam numerical method by the influence of Prandtl number, Eckert number, and thermal slip. The processing of NN-BLMT training, testing, and validation, is employed for various scenarios and cases to find and compare approximation solutions with reference results. For the fluidic system SF-HTM, convergence analysis based on mean square errors, histogram presentations, and statistical regression plots is considered for the proposed computing infrastructure’s performance in terms of NN-BLMT. Matching of the results for the fluid flow system SF-HTM based on proposed and reference results in terms of convergence up-to 10−07 to 10−03 proves the worth of proposed NN-BLMT.