NN Model Research Articles

IMPROVING CLUSTER ACCURACY IN TUITION FEES: A MULTILAYER PERCEPTRON NEURAL NETWORK AND RANDOM FOREST APPROACH

Manual classification of Single Tuition Fees (STF) has a high risk of misclassification due to the need for a more in-depth assessment of students' economic criteria. This research uses Artificial Neural Networks (ANN), specifically the Multilayer Perceptron (NN-MLP) model, to detect and correct errors in Single Tuition Fee (STF) classification. This study aims to apply the NN model to identify and correct classification errors in the STF clustering of State Islamic Religious Universities in Indonesia (PTKIN). This research was conducted using exploratory methods and quantitative approaches involving a population of PTKIN students throughout Indonesia. A sample of 282 respondents was selected using a simple random sampling method. The results showed that NN-MLP is an effective tool for identifying and correcting misclassification in determining PTKIN tuition fees with significantly improved classification accuracy characterized by an accuracy value of 71.28% and MSE of 0.287; this model can be used as a basis for developing information systems that are fairer and more accurate in managing tuition fees in higher education. This research also proves that the NN method is superior to traditional statistical methods and simple machine learning in handling complex and diverse data. In addition, the Random Forest model successfully identified the most influential input variables in STF classification. Father's occupation, mother's occupation, number of dependents, and utility bills such as water and electricity significantly contributed to the STF classification. In contrast, variables such as vehicle facilities showed a lower contribution.

Mixed-effects neural network modelling to predict longitudinal trends in fasting plasma glucose

BackgroundAccurate fasting plasma glucose (FPG) trend prediction is important for management and treatment of patients with type 2 diabetes mellitus (T2DM), a globally prevalent chronic disease. (Generalised) linear mixed-effects (LME) models and machine learning (ML) are commonly used to analyse longitudinal data; however, the former is insufficient for dealing with complex, nonlinear data, whereas with the latter, random effects are ignored. The aim of this study was to develop LME, back propagation neural network (BPNN), and mixed-effects NN models that combine the 2 to predict FPG levels.MethodsMonitoring data from 779 patients with T2DM from a multicentre, prospective study from the shared platform Figshare repository were divided 80/20 into training/test sets. The first 10 important features were modelled via random forest (RF) screening. First, an LME model was built to model interindividual differences, analyse the factors affecting FPG levels, compare the AIC and BIC values to screen the optimal model, and predict FPG levels. Second, multiple BPNN models were constructed via different variable sets to screen the optimal BPNN. Finally, an LME/BPNN combined model, named LMENN, was constructed via stacking integration. A 10-fold cross-validation cycle was performed using the training set to build the model and evaluate its performance, and then the final model was evaluated on the test set.ResultsThe top 10 variables screened by RF were HOMA-β, HbA1c, HOMA–IR, urinary sugar, insulin, BMI, waist circumference, weight, age, and group. The best-fitting random-intercept mixed-effects (lm22) model showed that each patient’s baseline glucose levels influenced subsequent glucose measurements, but the trend over time was consistent. The LMENN model combines the strengths of LME and BPNN and accounts for random effects. The RMSE of the LMENN model ranges were 0.447–0.471 (training set), 0.525–0.552 (validation set), and 0.511–0.565 (test set). It improves the prediction performance of the single LME and BPNN models and shows some advantages in predicting FPG levels.ConclusionsThe LMENN model built by integrating LME and BPNN has several potential applications in analysing longitudinal FPG monitoring data. This study provides new ideas and methods for further research in the field of blood glucose prediction.

Short to medium-term forecasting of fishing ground distribution based on deep learning model

Most of the fishing ground research focuses on real-time predictions and lacks continuous forecasting for future changes over a certain period. Traditional models characterized by large, single temporal scales lack effectiveness in accounting for the autocorrelation of environmental factors. Deep learning has demonstrated superior performance and promising development prospects due to its accurate and efficient ability to mine nonlinear information in the era of big data. Therefore, we take Ommastrephes bartramii as an example and constructed 28 different temporal scales and lead periods cases based on U-Net and compared them with GAM, NN and ConvLSTM model results. The input factors of this model are sea surface temperature (SST) and the output factors are the center fishing ground data (1998-2019). The results reveal that the optimal temporal scale and lead period for this model is 15 days and 4 on U-Net. The SST fluctuation information between different lead periods of environmental field and the coupling degree in fishing grounds may be essential factors affecting the model performance differences. It enhances marine fisheries understanding from artificial intelligence.

Forecasting Models for Time and Cost Performance Predicting of Infrastructural Projects

Abstract Developing artificial forecasting models to predict the cost index and time index is the primary objective of this research. The efficacy of residential property investment projects will be assessed through the implementation of models that incorporate Artificial Neural Networks and Multiple Linear Regression. Historical information of thirteen boundaries for twenty finished Private Property Venture Tasks were separated from the records of the Directorate of Lodging, then four models were created by utilizing Multiple Linear Regression strategy and Artificial Neural Networks method. The main model was Time Index model with test information that gave average accuracy equivalent to 78.48 % and delivered a high connection coefficient (R) equivalent to 95.3 % utilizing (MLR), the subsequent model was (CI) model with test information gave (AA) equivalents to 66.6 % and created (R) equivalents to 85.4 % utilizing (MLR), and the third model was (TI) model with test information gave (AA) equivalents to 90.31 %, and delivered a high (R) equivalents to 95.4 % utilizing (ANN), the fourth model was (CI) model with test information gave (AA) equivalents to 90.21 % and delivered a high (R) equivalents to 99.6 % utilizing (ANN). It was found that the NN model showed more exact assessment results than the MLR models. Thus, it was resolved that the NN model is generally appropriate for anticipating the time execution and cost execution of Residential Property Investment Projects.

A Stability Analysis of Neural Networks and Its Application to Tsunami Early Warning

AbstractNeural networks (NNs) enable precise modeling of complicated geophysical phenomena but can be sensitive to small input changes. In this work, we present a new method for analyzing this instability in NNs. We focus our analysis on adversarial examples, test‐time inputs with carefully crafted human‐imperceptible perturbations that expose the worst‐case instability in a model's predictions. Our stability analysis is based on a low‐rank expansion of NNs on a fixed input, and we apply our analysis to a NN model for tsunami early warning which takes geodetic measurements as the input and forecasts tsunami waveforms. The result is an improved description of local stability that explains adversarial examples generated by a standard gradient‐based algorithm, and allows the generation of other comparable examples. Our analysis can predict whether noise in the geodetic input will produce an unstable output, and identifies a potential approach to filtering the input that enable more robust forecasting.

Performance Analysis of Different Types of Nn Models for Target Recognition

The objective of this research is to conduct a comprehensive performance analysis of various types of neural network (NN) models for target recognition. Specifically, this study focuses on evaluating the effectiveness and efficiency of yolov8n, yolov8s, yolov8m models in target recognition tasks. Leveraging cutting-edge technologies such as OpenCV, the research is aimed at developing a robust methodology for assessing the performance of these NN models. Through meticulous analysis, this study aims to provide insights into the strengths and weaknesses of each model, facilitating informed decision-making for practical applications1. This paper presents the process of designing and conducting the performance analysis. The study discusses the implications of the findings for future developments in target recognition systems.

Deep learning modeling to assess surface temperature and greenery impact on emergency room access

Abstract The escalating land surface temperature is a key environmental indicator with potential repercussions on human health. An artificial NN model was built to predict the number of ER accesses as a dependent variable, to establish the predictive power and interactions of land surface temperature and normalized differential vegetation index(NDVI), which is used to measure urban greenery.A map of Taranto was built by vector files on the Apulia region website and processed using R Studio 2023.06.2.The vector file containing the mapping of the city of Taranto was processed to obtain a 200mx200m grid of the entire surface.The Local Health Authority of Taranto provided indications of the residences of ER accesses, and coordinates were merged to gridded vector file to obtain the average of ER accesses and the socio demographic features per single grid unit. NDVI and surface temperature were averaged for July 2023 using images from Landsat 8(30m) and merged to the vector providing an average value of temperature and NDVI index per single quadrant. The dataset was divided into two db according to an 80/20 ratio respectively for the training and test dataset.A NN model was built on dichotomous value (above/below median) for each quadrant (6513 units) with hyperparameter tuning technique using the accuracy to select the best model.A MOC model was built to reduce the probability of NN detection by at least 40% in a randomly selected subject.A MOC model was built considering sociodemographic variables as fixed features and an epsilon of 0 for 175 generations and an ICE curves population initialization strategy. NNET accuracy value was 0.849 and 0.841 in training and testing, suggesting a good fitting of the model.Relative change plots were adopted to show the amount and the direction of generated counterfactuals, to reduce grid unit prediction to be over median.LST decreasing and NDVI score increasing were associated with a reduction of classification probability of grid unit. Key messages • Rising land surface temperature is a critical environmental indicator with potential impacts on human health. • Artificial neural network models perform well in predicting the number of ER accesses with respect to land surface temperature and NDVI, which is used to measure urban greenery.

Speed and accuracy investigations of neural network algorithms for ionospheric modelling at an equatorial region

Neural networks are very efficient tools for modeling, including ionospheric modeling. The training algorithm is important for achieving the optimum performance of the trained network. This research is therefore meant to evaluate and compare the performances of ten neural network training algorithms based on their prediction accuracies, and the duration/times taken by each of the algorithms to establish the optimum result. The neural networks were trained using electron density measurements by Radio Occultation (RO) technique from the Constellation Observing System for Meteorology, Ionosphere, and Climate (COSMIC) satellites. Data for the period 2006 through 2021 was used. The networks were trained using about 2.9 million data points collected from the Kano region, Nigeria (5-degree rectangular region around geographic: 12.00° N, 8.59° E) after performing data quality control. The training algorithms considered in the work include: Bayesian Regularization (BR); Broyden-Fletcher-Goldfarb-Shanno Quasi-Newton (BFG); Conjugate Gradient with Powell/Beale (CGB); Fletcher-Reeves Conjugate Gradient (CGF); Gradient descent with momentum and adaptive learning rate (GDX); Levenberg Marquardt (LM); One Step Secant (OSS); Polak-Ribiére Conjugate Gradient (CGP); Resilient Backpropagation (RP); and Scaled Conjugate Gradient (SCG). The results showed that the BR and the LM algorithms gave the best performances in minimizing the errors of prediction (the mean RMSEs are respectively 112 and 114 ×103 electrons/cm3), but the RP algorithm, which came third in terms of accuracy, was significantly faster than both the LM and BR algorithms. The worst-performing algorithm in terms of accuracy was the GDX algorithm, although it was the fastest algorithm. The BFG algorithm was the worst-performing algorithm in terms of a combination of speed and accuracy. The developed neural network model was validated using ionosonde electron density measurements obtained from Ilorin, Nigeria (geographic: 8.5° N, 4.5° E; geomagnetic: 1.8° S). A comparison of the neural network, the NeQuick, and the IRI model predictions relative to the ionosonde measurements indicate that the neural network model was the best-performing model; the NN model predictions minimized the mean absolute errors (MAEs) in ∼44% of 399 ionosonde profiles investigated, the IRI model did so in ∼32%, and the NeQuick did so in ∼24%. The MAEs of the NeQuick however exhibited the best (least) variance. In overall, the NN model gave the least (best) mean of the MAEs (∼73 × 103 cm−3), compared to ∼82 × 103 cm−3 given by both the NeQuick and the IRI models, further supporting the idea that neural networks are excellent for present-day ionospheric modeling.

A hybrid deep learning-based short-term forecast model for ionospheric foF2 in East Asia region

The short-term forecast of the critical frequency of the ionospheric F2 layer (foF2) is particularly challenging due to its nonlinear and non-stationary characteristics. To improve the short-term forecast accuracy of foF2, we propose a short-term hybrid deep learning model that integrates the improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) and long short-term memory (LSTM) algorithms. The proposed model leverages the ICEEMDAN decomposition algorithm to expand foF2 time series data into multi-dimensional space and utilizes the LSTM algorithm to preserve long-term data information, capturing detailed changes in foF2 parameters and obtaining internal information about foF2. The 2014 and 2017 foF2 data from four observation stations located in Mohe, Beijing, Jeju, and Kokubunji in the mid-latitudes of East Asia were utilized for modeling and 1-hour forecast analysis. The forecast values of the proposed model align closely with the measured data, exhibiting minimal fluctuation in forecast error and effectively tracking the changing trend of foF2. The average root mean square error (RMSE) across the four stations in 2014 is 0.40 MHz, with a relative RMSE (RRMSE) of 7.11 % and a coefficient of determination (R2) of 0.98. The average RMSE in 2017 is 0.43 MHz, RRMSE is 7.68 %, and R2 is 0.92. The proposed model reveals significant improvements in the forecast accuracy of the comparison model compared with the CCIR, URSI, NN, and LSTM models. Especially compared with the CCIR and URSI models in the International Reference Ionosphere (IRI), demonstrating superior real-time tracking capability for foF2 and effectively addressing the challenge of poor short-term forecast accuracy observed in the IRI model.

Experimental Analysis and Hybrid-Optimization of Micro-ECDM Process Parameters to Enhance Micro-Machining Performances of Silica by Gaussian-Quantum-PSO

Nowadays, hybridization of different algorithms for the optimization of non-conventional machining processes tries to accomplish better results. The paper consists of experimental evolutionary-particle Swarm Optimization (PSO), Quantum-PSO and Gaussian Quantum Particle Swarm Optimization (G-QPSO)-based ANN modeling and comparative investigation on performances such as material removal rate (MRR), machining depth (MD), roughness of surface and overcut (OC) for machining of silica by ECDM process using mixed electrolyte. The paper also shows the co-efficient of NN models for different machining criteria and G-QPSO and also the comparative study of MD, roughness (SR), overcut (OC) as well as MRR using different algorithms and convergence test for fitness of experimental results also propounded to achieve cross-validation of models and multi-response optimal results for micro-machining of Silica by ECDM using PSO, QPSO and GQPSO. It is found that Gaussian Quantum Particle Swarm Optimization (G-QPSO)-ANN is more efficient for ECDM and achieves optimal results at 55-volt, pulse on time 52.3 s, inter-electrode gap (IEG) 30 mm, duty ratio 0.475 and electrolytic concentration 30 (wt.%).

Estimating the effect of sand-roughened bed on hydraulic jump characteristics using heuristic models

The hydraulic jump occurs in open channels and downstream of hydraulic structures because these features change the flow regime from super-critical to sub-critical. To prevent downstream channel erosion and guarantee the construction of a hydraulic jump, a stilling basin is utilized. The effect of sand-roughened beds on hydraulic jump characteristics was investigated using an experimental study. After that, new machine learning models including Emotional neural network (EANN), Hammerstein Weiner model (HWM), Elman neural networks ENN, and Extreme learning neural networks (ELNN) were adopted to estimate the ratio of hydraulic jump length to initial water depth. Different evaluation criteria such as determination Coefficient (R2), Mean Square Error (MSE), Correlation Coefficient (R), and Rout Mean Square Error (RMSE), scatter plots, time series comparison, and Spiral plots have been used to evaluate the models' performance. The results revealed that the Elman NN model outperformed the other models in estimating the ratio of hydraulic jump length to the initial depth (Lj/y1), while the HWM model has a weaker performance than other models in estimating Lj/y1.

Classification and prediction of drought and salinity stress tolerance in barley using GenPhenML

Genetic and agronomic advances consistently lead to an annual increase in global barley yield. Since abiotic stresses (physical environmental factors that negatively affect plant growth) reduce barley yield, it is necessary to predict barley resistance. Artificial intelligence and machine learning (ML) models are new and powerful tools for predicting product resilience. Considering the research gap in the use of molecular markers in predicting abiotic stresses, this paper introduces a new approach called GenPhenML that combines molecular markers and phenotypic traits to predict the resistance of barley genotypes to drought and salinity stresses by ML models. GenPhenML uses feature selection algorithms to determine the most important molecular markers. It then identifies the best model that predicts atmospheric resistance with lower MAE, RMSE, and higher R2. The results showed that GenPhenML with a neural network model predicted the salinity stress resistance score with MAE, RMSE and R2 values of 0.1206, 0.0308 and 0.9995, respectively. Also, the NN model predicted drought stress scores with MAE, RMSE and R2 values of 0.0727, 0.0105 and 0.9999, respectively. The GenPhenML approach was also used to classify barley genotypes as resistant and stress-sensitive. The results showed that the accuracy, accuracy and F1 score of the proposed approach for salinity and drought stress classification were higher than 97%.

NN MODELLING OF THE ENERGY CONSUMPTION IN THE WELDING PROCESS

Energy consumption is different for various technological processes used in manufacturing (cutting, plastic deformation, sintering, welding, etc.) and nowadays is becoming a more and more important issue to be considered when planning certain manufacturing processes. Related to this, when releasing a quotation for a givenproduct request coming for the market, it is important to have a tool enabling to evaluate the estimated energy consumption that will be required. This paper proposes a method to predict the consumption of energy in the welding process depending on the same main parameters of the process. The method is based on causal identification and on the NN modelling technique. Method application is sampled in the case of a real database including information concerning pipes welding. The method proves to be fast and delivers results of reasonable precision.

Enhanced Emotion Detection Model Using Dusky Canidae Optimization-Based Deep CNN Classifier

Emotion detection involves distinguishing the complex nature of emotions, cross-cultural variations, and ethical considerations. The proposed model addresses cross-cultural issues through diverse training and ensures ethical data usage through advanced machine-learning techniques to enhance accuracy. The research aims to develop a Dusky Canidae Optimization-based Deep Convolutional Neural Network model (DCO opt deep CNN) for emotion recognition in audio-visual data, benefiting various advancements. During preprocessing, audio and video are separated to enhance sound quality and eliminate data redundancy while retaining vital information. Features are then extracted and employed for training and testing using the automation and effectiveness of Deep CNNs. The concatenated features are optimized using a hybrid approach that combines Dusky Hawk and Canidae Wolf Optimization (CWO), offering simplicity and interpretability with optimal parameter settings. The proposed DCO model effectively classifies emotions within the data. Evaluating the database, the DCO model exhibits a performance enhancement of 1.28%, 2.91%, and 1.28% in accuracy, sensitivity, and specificity with respect to the existing learner memorizing optimization-based deep NN model.

Comparative Analysis of Supervised Classification Algorithms for Residential Water End Uses

AbstractWater sustainability in the built environment requires an accurate estimation of residential water end uses (e.g., showers, toilets, faucets, etc.). In this study, we evaluate the performance of four models (Random Forest, RF; Support Vector Machines, SVM; Logistic Regression, Log‐reg; and Neural Networks, NN) for residential water end‐use classification using actual (measured) and synthetic labeled data sets. We generated synthetic labeled data using Conditional Tabular Generative Adversarial Networks. We then utilized grid search to train each model on their respective optimized hyperparameters. The RF model exhibited the best model performance overall, while the Log‐reg model had the shortest execution times under different balanced and imbalanced (based on number of events per class) synthetic data scenarios, demonstrating a computationally efficient alternative for RF for specific end uses. The NN model exhibited high performance with the tradeoff of longer execution times compared to the other classification models. In the balanced data set scenario, all models achieved closely aligned F1‐scores, ranging from 0.83 to 0.90. However, when faced with imbalanced data reflective of actual conditions, both the SVM and Log‐reg models showed inferior performance compared to the RF and NN models. Overall, we concluded that decision tree‐based models emerge as the optimal choice for classification tasks in the context of water end‐use data. Our study advances residential smart water metering systems through creating synthetic labeled end‐use data and providing insight into the strengths and weaknesses of various supervised machine learning classifiers for end‐use identification.

Predictive Modelling of Traffic Flow with Deep Learning Techniques

This project delves into predictive modeling for traffic flow using deep learning techniques, focusing on the Metro Interstate dataset. Traffic Flow Prediction (TFP) is crucial for Intelligent Transport Systems (ITS), optimizing vehicle movement, reducing congestion, and improving route efficiency. Leveraging advancements in Artificial Intelligence (AI), Machine Learning (ML), Deep Learning (DL), and Big Data, our study explores various techniques and models for TFP. We highlight DL models' advantages over traditional ML methods, propelled by the wealth of real-time traffic data fostered by smart cities, presenting opportunities to craft robust predictive models. The core of our project revolves around developing a multi-step Recurrent Neural Network (RNN) with Long Short-Term Memory (LSTM) architecture. Our model forecasts traffic volume between Minneapolis and St. Paul, Minnesota, predicting volume two hours into the future based on a six-hour historical window. We explore DL algorithms' efficacy, including LSTM and Gated Recurrent Unit (GRU), in mitigating challenges like the vanishing gradient problem common in RNNs. Our analysis compares various NN models, emphasizing the importance of data availability for training and fine-tuning ML/DL models, with the Metro Interstate dataset serving as a crucial asset for comprehensive traffic flow analysis and model development.

Practical application of quantum neural network to materials informatics

Quantum neural network (QNN) models have received increasing attention owing to their strong expressibility and resistance to overfitting. It is particularly useful when the size of the training data is small, making it a good fit for materials informatics (MI) problems. However, there are only a few examples of the application of QNN to multivariate regression models, and little is known about how these models are constructed. This study aims to construct a QNN model to predict the melting points of metal oxides as an example of a multivariate regression task for the MI problem. Different architectures (encoding methods and entangler arrangements) are explored to create an effective QNN model. Shallow-depth ansatzs could achieve sufficient expressibility using sufficiently entangled circuits. The “linear” entangler was adequate for providing the necessary entanglement. The expressibility of the QNN model could be further improved by increasing the circuit width. The generalization performance could also be improved, outperforming the classical NN model. No overfitting was observed in the QNN models with a well-designed encoder. These findings suggest that QNN can be a useful tool for MI.

A horizontal vibration prediction method of high-speed elevator based on transferred-digital twin model

The vibrations of high-speed-elevators (HSEs) significantly impact the comfort of elevator rides, with horizontal vibrations (HV) being the most sensitive to bodies. Customized elevator products are characterized by high-speed capabilities, customization options, and complex operational conditions. These characteristics often lead to challenges such as difficulties in predicting HV, low prediction accuracy, and limited reusability of prediction models and data. In order to mitigate HV in elevators, its crucial to implement HV prediction during the design phase. To address this need, this paper proposes a method for predicting HSEHV based on a Transferred Digital Twin (TLDT) model. The HSE’s design is physically extracted, and an elevator Digital Twin (DT) model with a digital layer is constructed. This model includes product geometry models, dynamics models, and simulation models. A transfer learning (TL) model is developed to integrate simulation data generated by the digital layer with measurement data acquired from the data interface layer, resulting in high-capacity and high-fidelity DT data. Using the support vector regression (SVR) method, a high-dimensional nonlinear HSEHV prediction model is established, and is trained and optimized using DT data to achieve data-driven HSEHV prediction. The validity of this approach is confirmed using measured data from an elevator test tower. Results indicate that the SVR-based TLDT model achieves the highest accuracy and the lowest absolute error values for both peak-to-peak and A95 vibration acceleration when compared to other models such as NN, KNN, decision tree, and lasso models. Furthermore, by incorporating DT data, the model accuracies improve by 16.9%, 14.5%, 8.6%, 14.1%, and 8.6%, and 12.3%, 12.4%, 19.7%, 7.6%, and 8.7%, respectively, for the mentioned models. Finally, the HV of KLK2 elevator car system is predicted. When compared with simulation analysis and the conventional SVR prediction model, obtained results through the proposed method closely align with the measured values.

Multiparameter gas sensor based on UV-SFR combined with wavelet transform-neural network at ppb level

Abnormal concentrations of nitric oxide (NO), ammonia (NH3), carbon disulfide (CS2) and sulphur dioxide (SO2) in exhaled breath are closely associated with diseases in specific organs of the body. Meanwhile, the above gases are also common pollutants in the atmosphere, and effective online detection of them plays a key role in disease prevention and environmental pollution prevention. This study reports a multiparameter gas sensor system based on the combination of ultraviolet segmented fitting reduction (UV-SFR) and wavelet transform-neural network (WT-NN) model for simultaneous online detection of the concentrations of NO, NH3, CS2 and SO2. Firstly, to avoid the occurrence of spectral baseline drift and absorption peak intensity reduction, a UV-SFR method is proposed for the discrete single peak type absorption spectra contained in the UV spectra. Our method can remove the slow change absorption induced by other factors, and ultimately obtain the differential absorption spectra of the gas mixtures. Then, the WT is used to denoise the differential absorption spectra of the gas mixtures. Next, for the presence of severe spectral line overlap in gas mixtures, the NN model is built based on the denoised spectral data. Finally, transfer learning is introduced into NN to improve the detection accuracy. The experimental results show that the sensor system can reliably achieve simultaneous online detection of NO, NH3, CS2, and SO2 at ppb level, and the average absolute errors of their measurements are 0.96 ppb, 1.27 ppb, 0.76 ppb and 0.97 ppb, with measurement accuracies of 3.5 %, 0.8 %, 0.5 % and 0.9 %, respectively.

Prediction and optimizing of methylene blue sequestration to activated charcoal/magnetic nanocomposites using artificial neutral network and response surface methodology

Green synthesized magnetic nanoparticles (MNPs) linked with activated charcoal (AC) (AC/Fe3O4 NCs) were exploited for methylene blue (MB) confiscation in this study. The AC/Fe3O4 NCs produced were characterized using TEM, FTIR, UV/Vis and XRD spectrometry. The Response-Surface-Methodology (RSM) was utilized to improve the experimental data for the MB sorption to AC/Fe3O4 NCs, with 20 experimental runs implemented through a central composite design (CCD) to assess the effect of sorption factors-initial MB concentration, pH and sorbent dosage effects on the response (removal-effectiveness). The quadratic model was discovered to ideally describe the sorption process, with an R2 value of 0.9857. The theoretical prediction of the experimental data using the Artificial-Neural-Network (ANN) model showed that the Levenberg-Marquardt (LM) had a better performance criterion. Comparison between the modelled experimental and predicted data showed also that the LM algorithm had a high R2 of 0.9922, which showed NN model applicability for defining the sorption of MB to AC/Fe3O4 NCs with practical precision. The results of the non-linear fitting (NLF) of both isotherm and kinetic models, showed that the sorption of MB to AC/Fe3O4 NCs was perfectly described using the pseudo-second-order (PSOM) and Freundlich (FRHM) models. The estimated optimum sorption capacity was 455 mg g−1. Thermodynamically, the sorption of MB to AC/Fe3O4 NCs was shown to be non-spontaneous and endothermic.