Feed-forward Back-propagation Artificial Neural Network Research Articles

Application of Multiple Linear Regression and Artificial Neural Networks for the Prediction of the Packing and Capsule Filling Performance of Coated and Plain Pellets Differing in Density and Size.

Plain or coated pellets of different densities 1.45, 2.53, and 3.61 g/cc in two size ranges, small (380–550 μm) and large (700–1200 μm) (stereoscope/image analysis), were prepared according to experimental design using extrusion/spheronization. Multiple linear regression (MLR) and artificial neural networks (ANNs) were used to predict packing indices and capsule filling performance from the “apparent” pellet density (helium pycnometry). The dynamic packing of the pellets in tapped volumetric glass cylinders was evaluated using Kawakita’s parameter a and the angle of internal flow θ. The capsule filling was evaluated as maximum fill weight (CFW) and fill weight variation (FWV) using a semi-automatic machine that simulated filling with vibrating plate systems. The pellet density influenced the packing parameters a and θ as the main effect and the CFW and FWV as statistical interactions with the coating. The pellet size and coating also displayed interacting effects on CFW, FWV, and θ. After coating, both small and large pellets behaved the same, demonstrating smooth filling and a low fill weight variation. Furthermore, none of the packing indices could predict the fill weight variation for the studied pellets, suggesting that the filling and packing of capsules with free-flowing pellets is influenced by details that were not accounted for in the tapping experiments. A prediction could be made by the application of MLR and ANNs. The former gave good predictions for the bulk/tap densities, θ, CFW, and FWV (R-squared of experimental vs. theoretical data >0.951). A comparison of the fitting models showed that a feed-forward backpropagation ANN model with six hidden units was superior to MLR in generalizing ability and prediction accuracy. The simplification of the ANN via magnitude-based pruning (MBP) and optimal brain damage (OBD), showed good data fitting, and therefore the derived ANN model can be simplified while maintaining predictability. These findings emphasize the importance of pellet density in the overall capsule filling process and the necessity to implement MLR/ANN into the development of pellet capsule filling operations.

Prediction of vestibular schwannoma recurrence using artificial neural network.

ObjectivesTo compare two statistical models, namely logistic regression and artificial neural network (ANN), in prediction of vestibular schwannoma (VS) recurrence.MethodsSeven hundred eighty‐nine patients with VS diagnosis completed an online survey. Potential predictors for recurrence were derived from univariate analysis by reaching the cut off P value of .05. Those nine potential predictors were years since treatment, surgeon's specialty, resection amount, and having incomplete eye closure, dry eye, double vision, facial pain, seizure, and voice/swallowing problem as a complication following treatment. Multivariate binary logistic regression model was compared with a four‐layer 9‐5‐10‐1 feedforward backpropagation ANN for prediction of recurrence.ResultsThe overall recurrence rate was 14.5%. Significant predictors of recurrence in the regression model were years since treatment and resection amount (both P < .001). The regression model did not show an acceptable performance (area under the curve [AUC] = 0.64; P = .27). The regression model's sensitivity and specificity were 44% and 69%, respectively and correctly classified 56% of cases. The ANN showed a superior performance compared to the regression model (AUC = 0.79; P = .001) with higher sensitivity (61%) and specificity (81%), and correctly classified 70% of cases.ConclusionThe constructed ANN model was superior to logistic regression in predicting patient‐answered VS recurrence in an anonymous survey with higher sensitivity and specificity. Since artificial intelligence tools such as neural networks can have higher predictive abilities compared to logistic regression models, continuous investigation into their utility as complementary clinical tools in predicting certain surgical outcomes is warranted.

ON-Line MRI Image Selection and Tumor Classification using Artificial Neural Network

When soft tissue planning is important, usually, the Magnetic Resonance Imaging (MRI) is a medical imaging technique of selection. In this work, we show a modern method for automated diagnosis depending on a magnetic resonance images classification of the MRI. The presented technique has two main stages; features extraction and classification. We obtained the features corresponding to MRI images implementing Discrete Wavelet Transformation (DWT), inverse and forward, and textural properties, like rotation invariant texture features based on Gabor filtering, and evaluate the meaning of every property in the classification. The classifier is according to Feed Forward Back Propagation Artificial Neural Network (FP-ANN) in the classification stage. The properties thereafter derived to be implemented to teach a neural network based binary classifier that will be automatically able to conclude whether the image is that of a pathological, suffering from brain lesion, or a normal brain. The proposed algorithm obtained the sensitivity of 97.50%, specificity of 82.86% and accuracy of 94.3% for clinical Brain MRI database. This outcome proofs that the presented algorithm is robust and effective compared with other recent techniques.

DEVELOPING OPTIMAL ARTIFICIAL NEURAL NETWORK (ANN) TO PREDICT THE SPECIFIC HEAT OF WATER-BASED YTTRIUM OXIDE (Y2O3) NANOFLUID ACCORDING TO THE EXPERIMENTAL DATA AND PROPOSING NEW CORRELATION

In this study, the specific heat values of yttrium oxide-water nanofluid prepared in five different volumetric concentrations using Y2O3 nanoparticles were measured experimentally using the DTA method. Using the experimental results obtained, multilayer perceptron, feed-forward back-propagation artificial neural network with 15 neurons in its hidden layer was developed. Forty-two of the total 60 experimental data were used in the training phase, 12 in the validation phase, and 6 in the test phase. In addition, a new mathematical correlation has been proposed to calculate the specific heat values of yttrium oxide-water nanofluid. The artificial neural network has predicted the specific heat values of yttrium oxide-water nanofluid with an average error of -0.0007%. The error rate of the proposed new correlation was calculated as -0.011% on average.

Artificial Neural Networks Method for Prediction of Rainfall-Runoff Relation: Regional Practice

Rainfall and runoff relation is very important on efficient use of water resources and prevention of disasters. Nowadays, different methods of artificial intelligent techniques are applied to determine the rainfall-runoff relations. Artificial Neural Networks (ANN) is used for the present study. Also, Classical methods such as Multiple Linear Regression (MLR) are used. In this study, the data obtained from USA Waltham Massachusetts Stony Brook Reservoir basin was taken. 731 daily data of rainfall, runoff, and temperature were used to generate input data in the Multi Linear Regression and Feed-Forward Back-Propagation Artificial Neural Network models. The results obtained were compared with the actual results.

Feedforward Artificial Neural Network for Predicting Voltage Stability Indices in Power Systems

Several electricity failures associated with the voltage stability incident have appeared in a few countries. Nowadays, main concern towards voltage stability control and prediction is no longer crucial, however significant awareness is arising to sustain power system’s stability to conceal recurrence of major blackouts. Numerous types of line voltage stability indices (LVSI) being appointed to validate the weakest lines in IEEE 30-Bus test system. Besides that, LVSI is being forecasted by Feedforward Back Propagation Artificial Neural Network (FFBPNN) in order to recognize the voltage stability in IEEE 30-Bus test system. The calculated indices by using LVSI and forecasted indices by using FFBPNN are realistically applicable to discover the voltage collapse event in the system. The actual output for the VCPI(Power) in line 2-5 is 1.0459, while the predicted VCPI(Power) by using FFBPNN is 1.0459 with 3 seconds training time with 0% error percentage. Generally, the voltage collapse event has been successfully proven based on the capability of VCPI(Power). Therefore, necessary measures are capable to be performed by the power system operators to evade voltage collapse events occurred.

Using atmospheric inputs for Artificial Neural Networks to improve wind turbine power prediction

A robust machine learning methodology is used to generate a site-specific power-curve of a full-scale isolated wind turbine operating in an atmospheric boundary layer to drastically improve the power predictions, and, thus, the forecasting of the monthly energy production estimates. The study has important implication in measuring the financial feasibility of wind farms by improving the accuracy of monthly energy estimates. The significance of the study is that atmospheric stability and air-density are accounted in the power predictions of the wind turbine. Artificial Neural Networks (ANN) machine learning approach is used to generate multi-parameter input models to estimate the power produced by the wind turbine. The ANN model in this study uses Feed Forward Back Propagation (FFBP) algorithm. The power- and wind-data is obtained from a 2.5 MW wind turbine that has a Meteorological tower located 900 m Southwest of the wind turbine in Kirkwood, Iowa, USA. The study investigates the role of atmospheric boundary-layer metrics – Wind Speed, Density (a measure of stratification), Richardson Number, turbulence intensity, and wind shear as input parameters into the ANN model. The study investigates the influence of FFBP ANN hyper-parameters on the power prediction accuracy. Comparison of the FFBP ANN model to other power curve correction techniques demonstrated an improvement in the Mean Absolute Error (MAE) of 40% when compared to the density correction (the next closest). The five-parameter 4-layer FFBP ANN has an average energy production error of 0.4% for the nine months while the IEC this error is −3.7% and for the air density correction the error is −1.9%, respectively. Finally, the study determines the performance of the FFBP ANN model for different atmospheric stability regimes (Unstable, Stable, Strongly Stable, Strongly Unstable and Neutral) classified using two criterions - Richardson number and Turbulence intensity. The largest MAE occurs during the strongly stable regime of the atmospheric boundary layer for both criteria.

Classifying neural networks and methods of their illogical behaviour revealing

In this article a feedforward error backpropagation artificial neural network is investigated and the analysis of its illogical behaviour is presented. The problem of illogical behavior arises in various models of artificial neural networks. In the presented work a classifying artificial neural network (CANN) is considered and several learning algorithms were implemented and compared. CANN was designed for automatic differentiaition of cyanobacterial strains during environmental monitoring and some of trained networks demonstrated illogical behavior in further testing. Several original techniques were elaborated for estimation of the quality and accuracy of classification in addition to the traditional ones. Novel visualization methods were suggested for classification and generalization results representation.

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.

Determination of Mars Solar-Belt by Modeling of Solar Radiation Using Artificial Neural Networks

Missions to Mars need a power source, while, one of the most compatible sources for such a purpose is the photovoltaic system. Photovoltaic systems generate power based on the available energy from the Sun, and thus, solar radiation intensity at Mars should be known for design purposes. In this research, the feed-forward back-propagation artificial neural network is developed to predict solar radiation in terms of longitude, latitude, time of the day, temperature, altitude, pressure, amount of dust, and volume mixing ratio of water ice clouds. Data which are used to develop this model are obtained from the Mars Climate Database. The results of the developed method are accurate as compared with other methods whereas the correlation (R2) coefficient for the developed model is 0.97. The developed model then is used to predict mean solar radiation and mean temperature for every location on Mars and then the data are presented on Mars maps in order to determine the best location for harvesting energy from the Sun by photovoltaic systems. According to results, the solar radiation-temperature belt on Mars is found to be between latitudes 20 deg south and 15 deg north.

ATLASLang NMT: Arabic text language into Arabic sign language neural machine translation

ATLASLang is a machine translation system from Arabic text language into Arabic sign language (ArSL). The first version of the system (Brour and Benabbou, 2019) is based on two approaches: rule-based Interlingua and example-based approaches. It is a classical machine translation system, its limitations are the linguistic knowledge necessary to develop the rules, also many rules and exceptions required. In the last few years, a neural machine translation has achieved notable results, and several well-known companies including Google (Wu et al., 2016) and Systran (Crego et al., 2016) have starting to exploit it. In this paper, we present a new version of ATLASLang that is implemented using a feed-forward back-propagation Artificial Neural Network. In this version, we have started to translate simple sentences composed from a limited number of words since generally communication with deaf person uses short sentences. We have used the same ATLASLang MTS database of signs and we have utilized morphological characteristics to derive the maximum information from each word in the input of the neural translation system. The two versions of the system are compared using n-gram BLEU score (Papineni et al., 2002), and the results demonstrate that neuron network approaches outperform the other classical approaches.

Modeling the performance of Single-stage Nitrogen removal using Anammox and Partial nitritation (SNAP) process with backpropagation neural network and response surface methodology

Two novel feedforward backpropagation Artificial Neural Networks (ANN)-based-models (8:NH:1 and 7:NH:1) combined with Box-Behnken design of experiments methodology was proposed and developed to model NH4+ and Total Nitrogen (TN) removal within an upflow-sludge-bed (USB) reactor treating nitrogen-rich wastewater via Single-stage Nitrogen removal using Anammox and Partial nitritation (SNAP) process. ANN were developed by optimizing network architecture parameters via response surface methodology. Based on the goodness-of-fit standards, the proposed three-layered NH4+ and TN removal ANN-based-models trained with Levenberg-Marquardt-algorithm demonstrated high-performance as computations exhibited smaller deviations-(±2.1%) as well as satisfactory coefficient of determination (R2), fractional variance-(FV), and index of agreement-(IA) ranging 0.989–0.997, 0.003–0.031 and 0.993–0.998, respectively. The computational results affirmed that the ANN architecture which was optimized with response surface methodology enhanced the efficiency of the ANN-based-models. Furthermore, the overall performance of the developed ANN-based models revealed that modeling intricate biological systems (such as SNAP) using ANN-based models with the view to improve removal efficiencies, establish process control strategies and optimize performance is highly feasible. Microbial community analysis conducted with 16S rRNA high-throughput approach revealed that Candidatus Kuenenia was the most pronounced genera which accounted for 13.11% followed by Nitrosomonas-(6.23%) and Proteocatella-(3.1%), an indication that nitrogen removal pathway within the USB was mainly via partial-nitritation/anammox process.

Enhancing the Productivity of Wire Electrical Discharge Machining Toward Sustainable Production by using Artificial Neural Network Modelling

Sustainability plays an important role in manufacturing industries through economically-sound processes that able to minimize negative environmental impacts while having the social benefits. In this present study, the modeling of wire electrical discharge machining (WEDM) cutting process using an artificial neural network (ANN) for prediction has been carried out with a focus on sustainable production. The objective was to develop an ANN model for prediction of two sustainable measures which were material removal rate (as an economic aspect) and surface roughness (as a social aspect) of titanium alloy with ten input parameters. By concerning environmental pollution due to its intrinsic characteristics such as liquid wastes, the water-based dielectric fluid has been used in this study which represents an environmental aspect in sustainability. For this purpose, a feed-forward backpropagation ANN was developed and trained using the minimal experimental data. The other empirical modelling techniques (statistics based) are less in flexibility and prediction accuracy. The minimal, vague data and nonlinear complex input-output relationship make this ANN model simple and perfects method in the manufacturing environment as well as in this study. The results showed good agreement with the experimental data confirming the effectiveness of the ANN approach in the modeling of material removal rate and surface roughness of this cutting process.

Scour at bridge piers in uniform and armored beds under steady and unsteady flow conditions using ANN-APSO and ANN-GA algorithms

ABSTRACT Precise estimation of the local scour at bridge piers, considering flow unsteadiness and bed non-uniformity is crucial for economic and safe design of bridge piers. This study investigates scour at bridge piers in uniform and armored beds, under steady and unsteady flow conditions, applying feed-forward back-propagation (FFBP) artificial neural network (ANN) combined with evolutionary algorithms, using different data sets. ANN is well suited for complex pattern-sorting problems. However, their practical application is made difficult by the lack of a training algorithm, finding a global optimum weight in a relatively short time. Thus, evolutionary algorithms, including adaptive particle swarm optimization (APSO) and genetic algorithms (GA) are applied as exceptional tools for training and tuning the parameters of the proposed ANN models. Based on the available benchmark test data, the ANN-APSO model (with RMSE = 0.142 and 0.059 and R2 = 0.952 and 0.925 for steady and unsteady flow conditions, respectively) and ANN-GA model (with RMSE = 0.196 and 0.048 and R2 = 0.918 and 0.946 for steady and unsteady flow conditions, respectively) are shown to be superior to FFBP-ANN model (with RMSE = 0.214 and 0.106 and R2 = 0.891 and 0.838 for steady and unsteady flow conditions, respectively). Accordingly, the proposed ANN-APSO model is the best tool to evaluate the scour depth at bridge piers in uniform sediment beds under steady flow condition. It can be seen that the ANN-GA algorithm has the best fitness values compared to those of the ANN-APSO algorithm and the other reported approaches in uniform and armored sediment beds under unsteady flow condition. The developed models provide good approximations of the benefits of an FFBP-ANN combined with GA and APSO algorithms compared with a FFBP model, being superior in precision and time consumption.

Feasibility of simple machine learning approaches to support detection of non-glaucomatous visual fields in future automated glaucoma clinics.

To assess the performance of feed-forward back-propagation artificial neural networks (ANNs) in detecting field defects caused by pituitary disease from among a glaucomatous population. 24-2 Humphrey Visual Field reports were gathered from 121 pituitary patients and 907 glaucomatous patients. Optical character recognition was used to extract the threshold values from PDF reports. Left and right eye visual fields were coupled for each patient in an array to create bilateral field representations. ANNs were created to detect chiasmal field defects. We also assessed the ability of ANNs to identify a single pituitary field among 907 glaucomatous distractors. Mean field thresholds across all locations were lower for pituitary patients (20.3 dB, SD = 5.2 dB) than for glaucoma patients (24.4 dB, SD = 5.0 dB) indicating a greater degree of field loss (p < 0.0001) in the pituitary group. However, substantial overlap between the groups meant that mean bilateral field loss was not a reliable indicator of aetiology. Representative ANNs showed good performance in the discrimination task with sensitivity and specificity routinely above 95%. Where a single pituitary field was hidden among 907 glaucomatous fields, it had one of the five highest indexes of suspicion on 91% of 2420 ANNs. Traditional artificial neural networks perform well at detecting chiasmal field defects among a glaucoma cohort by inspecting bilateral field representations. Increasing automation of care means we will need robust methods of automatically diagnosing and managing disease. This work shows that machine learning can perform a useful role in diagnostic oversight in highly automated glaucoma clinics, enhancing patient safety.

Fabrication and modelling of the macro-mechanical properties of cross-ply laminated fibre-reinforced polymer composites using artificial neural network

The mechanical behaviour of fibre-reinforced polymer composites (FRPCs) is considered very complex due to many factors such as composition, material type, manufacturing process and end user applications. This article presents the mechanical properties and artificial neural network (ANN) modelling results of cross-ply laminated FRPCs. Twenty composite samples were fabricated by varying the number of layers of carbon fibre and glass fibre as reinforcement and polyphenylene sulphide and high-density polyethylene as matrix. Mechanical properties were measured in terms of flexural modulus, hardness, impact and transverse rupture strength. Multilayer feed-forward backpropagation ANN approach was used to predict the mechanical properties by using material type, composition and number of reinforcement and matrix layers as input variables. From 20 data patterns, 16 were used for network training and remaining 4 were used to test the models. Furthermore, trend analysis was also performed to understand the influence of inputs on developed models. It is evident from the ANN prediction results that there is good correlation between predicted and actual values within acceptable mean absolute error. The outcomes of this research will help to reduce cost and time by eliminating tedious composite property measurements and to fabricate tailored composites meeting application requirements.

Application of an artificial neural network model to predict the ultimate tensile strength of friction-welded titanium tubes

This paper presents an investigation to establish the link between friction welding process parameters and the ultimate tensile strength (UTS) of friction-welded titanium joints by application of an artificial neural network (ANN) technique. The experimental matrix is based on a central composite design with parameters varied at five levels. The UTS of the joints was modeled by the application of the response surface method (RSM). The joint UTS was also simulated by the application of a feed-forward back-propagation ANN with a single hidden layer composing of 20 neurons. The ANN was tested against and trained with the experimental data. The influence of the various parameters on the UTS was assessed by performing a sensitivity analysis. Lastly, the predictions of both the RSM model and the ANN were compared with one another. The results indicate that ANN is indeed a feasible technique for modeling and predicting the effect of process parameters on the UTS for friction welding of titanium tubes. When compared to RSM, ANN displayed a closer agreement with the data. In both cases, however, prediction errors were within 5%. Moreover, the link between the various process parameters and the UTS of the weld joints was also examined and commented upon.

Evaluation of static creep of FORTA-FI strengthened asphalt mixtures using experimental, statistical and feed-forward back-propagation ANN techniques

This paper investigates the effect of using different proportions (0%, 1%, 3%, 5%) of FORTA-FI fiber on asphalt mixtures static creep behavior under different compactive efforts (35 blows, 50 blows, 75 blows) and testing temperatures (25°C, 40°C, 55°C). The accumulated micro-strain was found to increase as temperature increases and stiffness modulus was found to decrease as temperature increases. These results are expected due to the viscoelastic nature of asphalt mixtures. Additionally, feed-forward back-propagation Artificial Neural Networks (ANN) was utilized in order to build a model that describes and predicts the relationship between deformation and stiffness modulus with multiple variables such as; Temperature (Temp), Fiber Content (FC) and Compactive Effort (CE). Evidently, a powerful predictive model was developed with Coefficient of Determination (R2) of 94%. The results using the SPSS statistical software of the two-way and three-way Analysis of Variance (ANOVA) tools showed that temperature and FORTA-FI content have significant interactions effects. Whereas, compactive effort effect on stiffness modulus was insignificant. Further, results obtained using Minitab statistical software confirmed the above findings.

Prediction and optimization studies for bioleaching of molybdenite concentrate using artificial neural networks and genetic algorithm

This paper presents the application of an artificial neural network (ANN) in order to predict the effects of operational parameters on the dissolution of Cu, Mo and Re from molybdenite concentrate through meso-acidophilic bioleaching. The initial pH, solid concentration, inoculum percent and time (days) were used as inputs to the network. The outputs of the models included the percent of Cu, Mo and Re recovered. The development and training of a feed-forward back-propagation artificial neural network (BPNN) was used to model and predict their recoveries. 105 sets of data were used to develop the neural network architecture and train it. To reach the network with highest generalizability, the space of neural networks with different hidden layers (one up to three hidden layers) and with the varying number of neurons each layer were searched. As a result, it was found that (4-5-5-2-1); (4-7-5-2-1) and (4-7-1-1-1) arrangements could give the most accurate prediction for Cu, Mo and Re extraction respectively. The regression analysis of the models tested gave a good correlation coefficient of 0.99968, 0.99617 and 0.99768 respectively for Cu, Mo and Re recoveries. The results demonstrated that ANN has a good potential to predict Cu, Mo and Re recoveries. Also, genetic algorithm (GA) was used to find out the optimum levels of parameters in the best models defined by ANN. The maximum recovery of Cu, Mo and Re on the 30th day were nearly 73%, 2.8% and 27.17% respectively.

Prediction of Bromate Removal in Drinking Water Using Artificial Neural Networks

ABSTRACTIn treatment of natural water resources, bromide transforms into carcinogenic bromate, especially during the ozonation process. Adsorption was used in the experimental part of this study to remove this harmful compound from drinking water. For this purpose, technically, HCl-, NaOH-, and NH3-modified activated carbons were used. Scanning Electron Microscopy (SEM) and Brunauer–Emmett–Teller (BET) analyses were carried out within the characterization study. Moreover, the effects of diameters and heights of adsorption columns, flowrate, and particle size of adsorbent were investigated on the removal amounts of bromate. Optimum conditions were obtained from the experiments, and regional/real samples were collected and analyzed. After the experiments, an artificial neural network (ANN) was used to predict bromate removal percentage by using the observed data. Within this context, a feed-forward back-propagation ANN was chosen in this study. Additionally, the transfer function was selected as tangent sigmoid and 3 neurons were used in the hidden layer. Particle size and amount of the activated carbon, height and diameter of the column, volumetric flowrate, and initial concentration were selected as the input variables. Bromate removal percentage was selected as the output. It was found that the model an R value of 0.988, RMSE value of 3.47 and mean absolute percentage error (MAPE) of 5.19% in the test phase.