Forward Back Propagation Artificial Neural Network Research Articles

Comparison of Swarm-based Metaheuristic and Gradient Descent-based Algorithms in Artificial Neural Network Training

This paper aims to compare the gradient descent-based algorithms under classical training model and swarm-based metaheuristic algorithms in feed forward backpropagation artificial neural network training. Batch weight and bias rule, Bayesian regularization, cyclical weight and bias rule and Levenberg-Marquardt algorithms are used as the classical gradient descent-based algorithms. In terms of the swarm-based metaheuristic algorithms, hunger games search, gray wolf optimizer, Archimedes optimization, and the Aquila optimizer are adopted. The Iris data set is used in this paper for the training. Mean square error, mean absolute error and determination coefficient are used as statistical measurement techniques to determine the effect of the network architecture and the adopted training algorithm. The metaheuristic algorithms are shown to have superior capability over the gradient descent-based algorithms in terms of artificial neural network training. In addition to their success in error rates, the classification capabilities of the metaheuristic algorithms are also observed to be in the range of 94%-97%. The hunger games search algorithm is also observed for its specific advantages amongst the metaheuristic algorithms as it maintains good performance in terms of classification ability and other statistical measurements.

Estimation of reference evapotranspiration using artificial neural network models for semi-arid region of Haryana

The study was conducted to evaluate performance of artificial neural network (ANN) models for estimating reference evapotranspiration (ET0) for semi-arid region of Haryana state. Ten years (2011-2020) daily weather data of maximum and minimum temperature, relative humidity, wind speed and sun shine hours was collected from the meteorological observatory at CCS HAU, Hisar. Multilayer perceptron feed forward back propagation ANN models were evaluated for different training algorithms (10), number of hidden layers (1-3) and number of neurons in hidden layers (1-30). Training algorithms compared in the study were heuristic techniques (GDA, GDX, RP), conjugate gradient (CGF, CGP, CGB, SCG), quasi-Newton (BFG, OSS) and Levenberg-Marquardt (LM). Results were compared against standard FAO Penman-Monteith method. The study revealed that best performance for ANN was found with LM algorithm in single layer of 13 neurons exhibiting RMSE, R, ME and RPD values 0.306, 0.986, 0.976 and 6.63, respectively. ANN models showed good performance in prediction of reference evapotranspiration.

Ku-Band SIW Filter with High Fractional Bandwidth Optimized Using Feed Forward Back Propagation ANN

An Artificial Neural Network (ANN) based Substrate Integrated Waveguide Band Pass Filter (SIW BPF) is proposed in this paper. A Feed Forward Back Propagation Neural Network (FF-BP NN) is utilized to optimize the filter dimensions. The Gradient Descent with Momentum and Adaptive Learning Rate algorithm is used to train the network at a learning rate of 0.01. The high pass characteristics of the SIW is converted into band pass by incorporating U slots on its top plane. Defected Ground Structure (DGS) is utilized in the bottom plane to improve the impedance matching. To validate, the prototype is fabricated using RT/Duroid 5880 and tested. The proposed filter has a center frequency at 14.68 GHz with a wide pass band from 12.92 to 16.43 GHz with a 3-dB Fractional Bandwidth (FBW) of 24%, return loss more than 20 dB and insertion loss of about 1.9 dB within the pass band. The filter has a small dimension of about 0.63 λ g 2 , where λ g is the guided wavelength at the center frequency. This filter offers wide passband, smaller size, low insertion loss, good return loss and it is useful in Ku-band satellite communication applications.

Elucidating the prediction capability of neural network model for estimation of crop water stress index of rice

ABSTRACT The crop water stress index is receiving significant attention these days, especially in arid and semiarid regions, for quantifying water stress and effective irrigation scheduling. Nowadays, machine learning techniques such as neural networks are being widely used to determine CWSI. In the present study, Self-Organizing Maps (SOM) and Feed Forward-Back Propagation Artificial Neural Networks (FF-BP-ANN) are compared while determining the CWSI of rice crop. Irrigation field experiments with varying degrees of irrigation were conducted at the irrigation field laboratory of the Indian Institute of Technology, Roorkee, during the growing season of the rice crop. The CWSI of rice was computed empirically by measuring key meteorological variables (relative humidity, air temperature, and canopy temperature). The empirically computed CWSI was compared with SOM and FF-BP-ANN predicted CWSI. For the lower CWSI baseline of rice, a linear relationship between air and canopy temperature difference (Tc-Ta) and, air vapour pressure deficit (AVPD) was developed, whereas air temperature plus 3°C was taken for the upper CWSI baselines. The performance of SOM and FF-BP-ANN were compared by computing Nash–Sutcliffe efficiency (NSE), index of agreement (d), root mean squared error (RMSE), and coefficient of correlation (R2). It is found that FF-BP-ANN (R2 = 0.97, NSE = 0.92, d = 0.95, RMSE = 0.006) performs better than SOM (R2 = 0.88, NSE = 0.87, d = 0.9, RMSE = 0.075).

Prediction of gamma ray spectrum for 22Na source by feed forward back propagation ANN model

The radiation has been used in a variety of different fields since its discovery and thus its measurement becomes vital in these industries. Different type detector may be used to measure gamma rays depends on the purposes of measurements. Gamma ray energy spectrum is an important to determine either elemental analysing of a sample or radiation shielding purposes. On the other hand, Artificial Neural Network (ANN) may be used to predict and analysing of gamma-ray spectrum. In this study, gamma ray spectrum from 22Na source detected in NaI (Tl) detector was estimated by ANN. There have been installed ten different ANN models to find the network structure that produces the best predictive value for the gamma ray spectrum NaI (Tl) Detector. Estimation study has been continued with the ANN model with be possessed of lowest error value. ANN model was created by using energy, distance and gamma-rays energy spectrum (called Io) values. In the ANN model developed using the feed forward back propagation algorithm, were used artificial neurons two in the input layer, ten in the hidden layer and one in the output layer. For the case of present work, the experimental data was used 70% for education, 20% for validation and 10% for testing. The estimated values obtained with the ANN model were compared with the experimental results and a good correlation has been found between them (R2 = 0.99).

Implementation of cognitive radio networks for optimum spectrum utilization through feed forward backpropagation artificial neural network

In wireless Radio Frequency (RF) networks, a huge amount of available bandwidth of the RF spectrum is wasted, because the users are not able to fully utilize the frequency band of the spectrum. The Cognitive Radio Networks provide an effective solution for optimum utilization of the frequency band of the RF spectrum. Cognitive Radio Networks consist of two types of users: primary users, which utilizes their allocated license frequency band of the RF spectrum and secondary users, which uses the license frequency band of the primary users when they are not using the RF spectrum. Primary users pre-empt the secondary users when they want to use the RF spectrum. To know when the spread spectrum is not occupied by the primary users, the secondary users must have some mechanism to sense the environment, learn the parameters and take a decision on the basis of it. The users of Cognitive Radio Networks enhance these capabilities by using Artificial Neural Networks (ANN). This paper discusses how the presence of primary users in the particular frequency band of the spectrum could be detected by the secondary users by using Feed Forward Back Propagation ANN. For this a CRN simulation model is designed and developed which uses a Feed forward back propagation ANN for optimum use of the spectrum.

Pull-in Phenomenon in the Electrostatically Micro-switch Suspended between Two Conductive Plates using the Artificial Neural Network

Artificial Neural Networks (ANN) are designed to evaluate the pull-in voltage of MEMS switches. The mathematical model of a micro-switch subjected to electrostatic force is preliminarily illustrated to get the relevant equations providing static deflection and pull-in voltage. Adopting the Step-by-Step Linearization Method together with a Galerkin-based reduced order model, numerical results in terms of pull-in voltage are obtained to be employed in the training process of ANN. Then, feed forward back propagation ANNs are designed and a learning process based on the Levenberg-Marquardt method is performed. The ability of designed neural networks to determine pull-in voltage have been compared with previous results presented in experimental and theoretical studies and it has been shown that the presented method has a good ability to approximate the threshold voltage of micro switch. Furthermore, the geometric and physical effect of the micro-switch on the pull-in voltage was also examined using these designed networks and relevant findings were provided.

Utilization of Cucurbit[6]uril as an effective adsorbent for the remediation of Phthalocyanine and Procion golden yellow dyes

The present study explores the utilization of the major product Cucurbit[6]uril (CB[6]) a macrocyclic host cavitand formed during synthesis of Cucurbit[ n ]urils, as a cost-effective adsorbent to remove C.I. Reactive orange 12 (RO-12) and C.I. Direct blue 199 (DB-199) dyes from effluent water. The effects of various parameters such as dye concentration, pH, adsorbent dosage, contact time, temperature and metal ions on the adsorption are thoroughly investigated. The performed isotherm calculations suggest the Langmuir to be the suitable isotherm model for adsorption of both dyes with maximum adsorption capacity 230 mg/g of RO-12 and 240 mg/g of DB-199 at room temperature. Kinetic studies reveal that the adsorption follows the pseudo-first-order kinetics for both the dyes. Thermodynamic calculations suggest that adsorption of both the dyes to be an enthalpy-driven and spontaneous. The analytical and density functional theory-based calculations were used to explain adsorption mechanism. These analyses revealed that adsorption process is dominated by the electrostatic interactions of the dye and CB[6]. In addition, A three-layer artificial neural network (ANN) model is used to predict the percentage of dye removal. A cascade Forward Back Propagation (CFBP) ANN was used with Levenberg–Marquardt algorithm (LMA), a linear transfer function (purelin) at output layer and a tangent sigmoid transfer function (tansig) in the hidden layer with 50 neurons. • CB[6] was utilized efficiently to remove Phthalocyanine and Reactive class of dye. • Adsorption behavior, isotherms, dynamics, and thermodynamics were studied. • Molecular mechanism of dyes adsorption on CB[6] was evaluated. • Adsorption mechanism were investigated by means of TGA, IR and DFT calculations. • Artificial neural network model was used to predict the percentage of dye removal.

Photocatalytic Removal of RhB by Ag and Mg Co-Doped ZnO Nanoparticles: Modeling of Operational Parameters Using ANN Based on RSM Data

In this study, using response surface methodology (RSM), we focused on optimizing Rhodamine B (RhB) removal via Ag, Mg co-doped ZnO nanoparticles. Statistical analysis and lack of fit indicated that the model was adequate for optimizing the process. The optimum conditions were 240 mg L–1 photocatalyst dosage, 5 mg L–1 RhB concentration and 10 min irradiation time. A mathematical equation was created between RhB removal percent and effective operational parameters. The results calculated through RSM equation were used for modeling the process via artificial neural networks (ANN). Operational parameters such as photocatalyst dosage, initial RhB concentration and irradiation time were selected as the input variables. A three-layered feed forward back propagation ANN model was developed to process modeling. Data used in RSM design was simulated for testing the accuracy of the model generated by ANN. The comparison of the ANN predicted data and the experimental data revealed that the proposed method is significantly more effective in terms of the process modeling. The results obtained via ANN modeling indicated that the photocatalyst dosage with the relative importance of 49.61% can be considered as the most effective parameter for RhB photocatalytic removal.

Prediction of gas production potential and hydrological properties of a methane hydrate reservoir using ANN-GA based framework

Methane hydrate reservoirs are a potential alternative energy source identified in the recent years. This work develops a methodology to predict the gas production potential and hydrological transport properties of methane hydrate reservoirs using artificial neural network coupled with genetic algorithm. Hydrate dissociation due to depressurisation in a lab scale core recovered from field excavation is modelled as multi-component, multi-phase, chemically reacting, non-isothermal flow in porous medium with effective transport properties. A predictive surrogate model, mimicking the full scale model along with reservoir heterogeneity is then developed using a feed forward back-propagation artificial neural network trained using the Levenberg-Marquardt algorithm coupled with Bayesian regularization. Predictions of gas production potential obtained from the surrogate model are then compared with those obtained from the full numerical simulations as well as measurements reported in literature for prescribed values of hydrological transport properties. It is then interfaced with genetic algorithm to solve an inverse estimation problem involving prediction of reservoir hydrological properties namely porosity, intrinsic permeability and initial hydrate saturation from temporally varying gas production. The effect of uncertainties associated with measurement of gas production on the predicted hydrological properties is investigated numerically and suitable bounds are proposed based on a statistical analysis.

Modelling of Kerf Width in Plasma Jet Metal Cutting Process using ANN Approach

In this paper Artificial Neural Network (ANN) model was developed for prediction of kerf width in plasma jet metal cutting process. Process parameters whose influence was analyzed are cutting height, cutting speed and arc current. An L18 (21x37) Taguchi orthogonal array experiment was conducted on aluminium sheet of 3 mm thickness. Using the experimental data a feed – forward backpropagation artificial neural network model was developed. After the prediction accuracy of the developed model was verified, the model was used to generate plots that show influence of process parameters and their interactions on analzyed kerf width and to get conlusions about process parameters values that lead to minimal kerf width.

Bridging Communication Gap Among People with Hearing Impairment: An Application of Image Processing and Artificial Neural Network

Before the present study, no sign language recognition system for the Nigeria indigenous sign language particularly Yoruba language has been developed. As a result, this research endeavors at introducing a Yoruba Sign Language recognition system using image processing and Artificial Neural Network (ANN).The proposed system (YSLRS) was implemented and tested. 600 images from 60 different signers were gathered. The images were acquired using vision based method, the different signers were asked to stand in front of a laptop’s camera make sign number from one to ten with their fingers in three different times and the images were stored in a folder. The image dataset was pre-processed for proper presentation for de-noising, segmentation and feature extraction. Thereafter, pattern recognition was done using feed forward back propagation ANN. The study revealed that Median filter with higher PSNR of 47.7 a lower MSE of 1.11, performed better than the Gaussian filter. Furthermore, the efficiency of the developed system was determined using mean square error and the best validation performance occurred at 25 epochs with a MSE of 0.004052, implying than ANN was able to adequately recognize the pattern of the Yoruba signs. Histogram was also used to determine the efficiency of the system, it can be seen that the histogram of the trained, tested and validated error bars were close to zero error, implying that the ANN and Receiver Operating Characteristic (ROC) was used to evaluate the performance of ANN in matching the features of the Yoruba Signs, which shows that ANN performed efficiently, having a high true positive rate and a minimum false positive rate. Finally, YSLRS developed in the study would reduce negative attitudes of victimizations suffered by the hearing-impaired individuals, by bridging communication gap among Nigerian PWD with hearing impairment.

An enhanced machine learning based approach for failures detection and diagnosis of PV systems

In this paper, a novel procedure for fault detection and diagnosis in the direct current (DC) side of PV system, based on probabilistic neural network (PNN) classifier, is proposed. The suggested procedure consists of four main stages: (i) PV module parameters extraction, (ii) PV array simulation and experimental validation (iii) elaboration of a relevant database of both healthy and faulty operations, and (iv) network construction, training and testing. In the first stage, the unknown electrical parameters of the one diode model (ODM) are accurately identified using the best-so-far ABC algorithm. Then, based on these parameters the PV array is simulated and experimentally validated by using a PSIM™/Matlab™ co-simulation. Finally, efficient fault detection and diagnosis procedure based on PNN classifier is implemented. Four operating cases were tested in a grid connected PV system of 9.54kWp: Healthy system, three modules short-circuited in one string, ten modules short-circuited in one string, and a string disconnected from the array. Moreover, the PNN method was compared, under real operating conditions, with the feed forward back-propagation Artificial Neural Network (ANN) classifiers method, for noiseless and noisy data to evaluate the suggested method’s accuracy and test its aptitude to support noisy data. The obtained results have demonstrated the high efficiency of the proposed method to detect and diagnose DC side anomalies for both noiseless and noisy data cases.

Experimental investigation and modeling of thermal conductivity of CuO–water/EG nanofluid by FFBP-ANN and multiple regressions

The purpose of this study is to predict the thermal conductivity of copper oxide (CuO) nanofluid by using feed forward backpropagation artificial neural network (FFBP-ANN). Thermal conductivity of CuO nanofluid is measured experimentally using transient hot-wire technique in temperature range of 20–60 °C and in volume fractions of 0.00125, 0.0025, 0.005 and 0.01% for neural network training and modeling. In addition, in order to evaluate accuracy of modeling in predicting the coefficient of nanofluid thermal conductivity, indices of root-mean-square error, coefficient of determination (R 2) and mean absolute percentage error have been used. FFBP-ANN with two input parameters (volume fraction and nanofluid temperature) and one output parameter (nanofluid thermal conductivity) in addition to two hidden layers and one outer layer which purelin, logsig and tansig functions are used was considered as the most optimum structure for modeling with neuron number of 4–10–1. In this study, among common methods of theoretical modeling of nanofluid thermal conductivity, theoretical method of Maxwell and also multivariate linear regression model was used for explaining the importance of modeling and predicting the results using neural network. According to this research, the results of indices and predictions show high accuracy and certainty of ANN modeling in comparison with empirical results and theoretical models.

Modeling and optimization of polymer enhanced ultrafiltration using hybrid neural-genetic algorithm based evolutionary approach

A stochastic genetic algorithm (GA) based strategy along with artificial neural network (ANN) was applied to optimize the retention of reactive red 120 (RR 120) dye from its aqueous solutions by way of polymer (polyethyleneimine (PEI)) enhanced ultrafiltration (PEUF). The optimal feed forward back propagation ANN (4-10-1) model network, trained initially through Levenberg–Marquardt (LM) algorithm, was suitably manoeuvred by the GA approach to predict the membrane performance index (PFI) response, evaluated as the product of dye rejection and permeation flux, for a randomly generated population of chromosomes. Each chromosome was constituted by four principal genes, namely, cross-flow rate, transmembrane pressure, polymer to dye ratio, and pH. The local exploitation capacity of the canonical GA was enhanced further by combining hill-climbing (HC) local search with the optimization levels of standard GA. The near-optimal and economically feasible factor levels were predicted by the hybrid ANN-GA-HC strategy, keeping PFI maximization and the constrained PEUF process dynamics in perspective; the optimal process factor settings experimentally yielded a pragmatic PFI of 143.8L/m2h, corresponding to high (99.9%) dye rejection, and a satisfactory permeation flux (144L/m2h).

Prediction of paddy drying kinetics: A comparative study between mathematical and artificial neural network modelling

The present study aimed at investigation of deep bed drying of rough rice kernels at various thin layers at different drying air temperatures and flow rates. A comparative study was performed between mathematical thin layer models and artificial neural networks to estimate the drying curves of rough rice. The suitability of nine mathematical models in simulating the drying kinetics was examined and the Midilli model was determined as the best approach for describing drying curves. Different feed forward-back propagation artificial neural networks were examined to predict the moisture content variations of the grains. The ANN with 4-18-18-1 topology, transfer function of hyperbolic tangent sigmoid and a Levenberg-Marquardt back propagation training algorithm provided the best results with the maximum correlation coefficient and the minimum mean square error values. Furthermore, it was revealed that ANN modeling had better performance in prediction of drying curves with lower root mean square error values.

Application of Artificial Neural Network for Predicting Shaft and Tip Resistances of Concrete Piles

<p class="MsoNormal" style="text-align: justify; line-height: 200%;">Axial bearing capacity (ABC) of piles is usually determined by static load test (SLT). However, conducting SLT is costly and time-consuming. High strain dynamic pile testing (HSDPT) which is provided by pile driving analyzer (PDA) is a more recent approach for predicting the ABC of piles. In comparison to SLT, PDA test is quick and economical. Implementing feed forward back-propagation artificial neural network (ANN) for solving geotechnical problems has recently gained attention mainly due to its ability in finding complex nonlinear relationships among different parameters. In this study, an ANN-based predictive model for estimating ABC of piles and its distribution is proposed. For network construction purpose, 36 PDA tests were performed on various concrete piles in different project sites. The PDA results, pile geometrical characteristics as well as soil investigation data were used for training the ANN models. Findings indicate the feasibility of ANN in predicting ultimate, shaft and tip bearing resistances of piles. The coefficients of determination, R², equal to 0.941, 0.936, and 0.951 for testing data reveal that the shaft, tip and ultimate bearing capacities of piles predicted by ANN-based model are in close agreement with those of HSDPT. By using sensitivity analysis, it was found that the length and area of the piles are dominant factors in the proposed predictive model.</p><p class="MsoNormal" style="text-align: justify; line-height: 200%;"> </p><p class="MsoNormal" style="text-align: justify; line-height: 200%;"><strong>Resumen</strong></p><p class="MsoNormal" style="text-align: justify; line-height: 200%;">La Capacidad Axial de Soporte (ABC, en inglés) de un pilote de construcción se determina usualmente a través de una Prueba de Carga Estática (SLT, inglés). Sin embargo, estas pruebas son costosas y demandan tiempo. La evaluación de las Dinámicas de Alto Esfuerzo de Pilotes (HSDPT, inglés), que la provee el programa de Análisis de Excavación (PDA, inglés), es una forma de aproximación más reciente para preveer la Capacidad Axial de Soporte. En comparación con la Prueba de Cargas Estática, la evaluación PDA es rápida y económica. La implementación de Redes Neuronales Arficiales (ANN, en inglés) que permita resolver problemas geotécnicos ha ganado atención recientemente debido a su posibilidad de hallar relaciones no lineales entre los diferentes parámetros. En este estudio se propone un modelo predictivo ANN para estimar la Capacidad Axial de Soporte de pilotes y su distribución. Para fines de una red de construcción se realizaron 36 pruebas PDA en pilotes de diferentes proyectos. Los resultados de los Análisis de Excavación, las características geométricas de los pilotes, al igual que los datos de investigación del suelo se utilizaron para probar los modelos ANN. Los resultados indican la viabilidad del modelo ANN en predecir la resistencia de los pilotes. Los coeficientes de correlación, R², que alcanzaron 0.941, 09.36 y 0.951 para la evaluación de los datos, revelan que la capacidad del pilotaje en el último rodamiento, en el cojinete del eje y en la punta que se predijeron con el modelo ANN concuerda con las establecidas a través del HSDPT. A través del análisis de respuesta se determinó que la longitud y el área de los pilotes son factores dominantes en el modelo predictivo propuesto.</p>

Implementation of Switched Beam Smart Antenna Using Artificial Neural Network

The capacity enhancement promised by switched beam smart antenna is a function of efficient selection of the active beam at each point in time. This article presents the use of artificial neural network (ANN) to improve the performance of switched beam smart antenna. The proposed method is based on feed forward backpropagation ANN. In this design, samples of the calibration of the footprint of the base station is mapped to the radiation pattern of Butler matrix (BM) base station antenna arrays. A $$4\times 4$$4×4 BM is designed to implement the developed technique and a digitizer applied to obtain the ANN training data. The results are compared with the existing numerical method and negative selection algorithm based on received signal strength indicator. It is shown that the proposed method exhibits a better switching performance.

Effective Earth Radius Factor Prediction and Mapping for Ondo State, South Western Nigeria

Accurate prediction and determination of the effective earth radius factor (k-factor) is critical for optimal performance in the design and planning of terrestrial line of sight communication links. In this work an artificial neural network (ANN) model is developed and used to predict k-factor values for four towns in Ondo State using satellite derived data. The towns are Okitipupa (6.5°N, 4.78°E), Ondo (7.11°N, 4.83°E), Akure (7.25°N, 5.2°E) and Ikare (7.52°N, 5.75°E). A feed forward back propagation ANN was implemented, thereafter trained, validated and tested using satellite derived data for the period from 1984 to 2002. Mean absolute error (MAE) was used to evaluate the performance of the ANN. The MAE values obtained were 0.0024, 0.0014, 0.002 and 0.003 for Okitipupa, Akure, Ikare and Ondo towns respectively. Contour map showing the predicted k-factor values interpolated over the map of Ondo state was plotted using Geographical Information System (GIS) techniques. The study concluded that ANN presents an effective means of predicting the average k-factor values over a geographical location.

Modeling and Analysis of Mechanical Properties of Aluminium Alloy (A413) Processed through Squeeze Casting Route Using Artificial Neural Network Model and Statistical Technique

Artificial Neural Network (ANN) approach was used for predicting and analyzing the mechanical properties of A413 aluminum alloy produced by squeeze casting route. The experiments are carried out with different controlled input variables such as squeeze pressure, die preheating temperature, and melt temperature as per Full Factorial Design (FFD). The accounted absolute process variables produce a casting with pore-free and ideal fine grain dendritic structure resulting in good mechanical properties such as hardness, ultimate tensile strength, and yield strength. As a primary objective, a feed forward back propagation ANN model has been developed with different architectures for ensuring the definiteness of the values. The developed model along with its predicted data was in good agreement with the experimental data, inferring the valuable performance of the optimal model. From the work it was ascertained that, for castings produced by squeeze casting route, the ANN is an alternative method for predicting the mechanical properties and appropriate results can be estimated rather than measured, thereby reducing the testing time and cost. As a secondary objective, quantitative and statistical analysis was performed in order to evaluate the effect of process parameters on the mechanical properties of the castings.