Artificial Neural Network Methodology Research Articles

Optimization of Video Stimuli Parameters in EMDR Therapy Using Artificial Neural Networks for Enhanced Treatment Efficacy

Eye Movement Desensitization and Reprocessing (EMDR) was recognized by the World Health Organization in 2013 as an evidence-based therapy for post-traumatic stress disorder (PTSD) and found to be effective for depression. Since then, EMDR has evolved into a personalized treatment focusing on stabilizing the physiological and psychological processes to alleviate symptoms of depression and stress. However, optimized parameters for video stimuli, such as speed (ssp), distance (d), and size (ssz), are not yet well defined in EMDR protocols. This study addresses this gap by employing an artificial neural network (ANN) methodology based on Francine Shapiro’s Adaptive Information Processing (AIP) model. The ANN was used to determine ideal values for video stimuli parameters, developing an integrated model to enhance EMDR outcomes. Of the 2860 ANN-modeled combinations, stimulus settings of 1.8 Hz speed, 70-pixel size, and 1440-pixel distance achieved the highest Predicted Effectiveness Score (PES) of 98.7%. An EMDR field test with electroencephalography (EEG) was conducted to assess the optimized video stimuli’s efficacy. Further, 16 participants, selected from a sample of 56 meeting CES-D depression criteria, were evaluated, and the top 50 PES values were selected for further analysis. EEG results indicated a 12.31% increase in effectiveness, showing a reduction in right frontal lobe beta waves. These findings highlight the technical advancements and therapeutic potential of the proposed ANN-optimized EMDR stimuli, demonstrating statistically significant improvements over traditional methods.

RSM modelling and yield prediction of base-catalyzed transesterification of pumpkin-seed oil to biodiesel: artificial neural network, kinetics, and thermodynamics

The study examined the extraction of bio-oil from pumpkin seed and compared the optimization of the production of Fatty acid ethyl ester (FAEE) via the transesterification process using Response Surface Methodology (RSM), and Artificial Neural Network (ANN). This research uniquely highlights the utilization of pumpkin seed oil, a non-edible and sustainable feedstock, and combines RSM and ANN methodologies to enhance the precision of biodiesel optimization. The transesterification experiment was conducted at 60 min reaction time under varying temperature ranges, catalyst weight, stirrer speed, and ethanol-oil molar ratio. The RSM optimized conditions for maximum production were determined to be a 1.3% catalyst concentration, 6:1 ethanol-to-oil molar ratio, 50 °C temperature, and 550 rpm stirrer speed, resulting in a 90% biodiesel yield. The statistical evaluation metrics confirmed the neural network predictions were compromised to 80% yield due to limitations such as insufficient data size and the inherent complexity of the ANN model. The biofuel produced satisfied ASTM specifications peculiar to sustainable environmental applications. The mechanistic parameters revealed variations in thermodynamic stability and feasibility across reaction orders (zero, pseudo-first, and pseudo-second), emphasizing the temperature-dependent effects of the transesterification kinetic pathway on yield, design, and efficiency.

Comparison of Different Machine Learning Methodologies for Predicting the Non-Specific Treatment Response in Placebo Controlled Major Depressive Disorder Clinical Trials.

Placebo effect represents a serious confounder for the assessment of treatment effect to the extent that it has become increasingly difficult to develop antidepressant medications appropriate for outperforming placebo. Treatment effect in randomized, placebo-controlled trials, is usually estimated by the mean baseline adjusted difference of treatment response in active and placebo arms and is function of treatment-specific and non-specific effects. The non-specific treatment effect varies subject by subject conditional to the individual propensity to respond to placebo. This effect is not estimable at an individual level using the conventional parallel-group study design, since each subject enrolled in the trial is assigned to receive either active treatment or placebo, but not both. The objective of this study was to conduct a comparative analysis of the machine learning methodologies to estimate the individual probability of a non-specific treatment effect. The estimated probability is expected to support novel methodological approaches for better controlling effect of excessively high placebo response. At this purpose, six machine learning methodologies (gradient boosting machine, lasso regression, logistic regression, support vector machines, k-nearest neighbors, and random forests) were compared to the multilayer perceptrons artificial neural network (ANN) methodology for predicting the probability of individual non-specific treatment response. ANN achieved the highest overall accuracy among all methods tested. A fivefold cross-validation was used to assess performances and risks of overfitting of the ANN model. The analysis conducted without subjects with non-specific effect indicated a significant increase of signal detection with significant increase in effect size.

Comparative analysis of coconut husk and laminated plastic packaging co-pyrolysis: Response surface methodology and artificial neural network approach

The co-pyrolysis of plastic with biomass offers a promising solution to mitigate environmental health concerns associated with plastic waste. This technique has shown effectiveness in addressing waste issues by generating fuel oil from plastic and biomass waste. Many researchers have utilized pyrolysis technology, employing various optimization techniques to produce significant amounts of pyrolytic oil. This study aims to compare predictions of the percentage mass oil yield using an artificial neural network (ANN) and response surface methodology (RSM). Before the pyrolysis process, a 3k factorial Box-Behnken Design was employed to determine the number of experiments required. In RSM, a 2D contour plot illustrated the correlation of each parameter with the percentage oil yield. ANOVA analysis revealed the significance of the produced quadratic mathematical model, with a p-value below 0.05. Through the ANN modeling, the temperature, particle size, and the percentage of LPP were employed as the input, while two neurons were employed in 1 hidden layer. The resulting percentage oil yields were calculated, indicating a significant influence from temperature and the percentage of laminated plastic packaging. Simulation results from the ANN demonstrated strong agreement with a correlation coefficient of 99.5%, surpassing the 90.71% correlation coefficient observed in RSM. This underscores the advantage of using ANN for predictive modeling.

Development of Comprehensive Predictive Models for Evaluating Böhme Abrasion Value (BAV) of Dimension Stones Using Non-Destructive Testing Methods

Due to the global demand for dimension stones, fast and reliable evaluation tools are essential for assessing the quality of dimension stones. For this reason, this study aims to develop comprehensive tools for estimating the abrasion resistance of various dimension stones from Turkey. Non-destructive rock properties, including dry density (ρd), water absorption by weight (wa), and pulse wave velocity (Vp), were determined to build a comprehensive database for soft computing analyses. Three predictive models were established using multivariate adaptive regression spline (MARS), M5P, and artificial neural networks (ANN) methodologies. The performance of the models was assessed through scatter plots and statistical indicators, showing that the ANN-based model outperforms those based on M5P and MARS. The applicability of the models was further validated with independent data from the existing literature, confirming that all models are suitable for estimating varying Böhme abrasion values (BAVs). A MATLAB-based software tool, called Böhme abrasion calculator (v1.00), was also developed, allowing users to estimate BAV values by inputting adopted non-destructive rock properties. This tool is available upon request, supporting the dimension stone industry and fostering future research in this field.

Application of eco-friendly material as an inexpensive adsorbent for methyl violet dye removal: experimental, response surface methodology and statistical physics

Methyl Violet (MV) removal from aqueous solutions is studied using an Algerian Bentonite sample as a low-cost adsorbent. SEM-EDX, X-ray diffraction, and chemical composition characterized the adsorbent material. The modeling and optimization study of MV adsorption using artificial neural network (ANN) and response surface methodology (RSM) were also examined. The effects of pH, contact time, dye concentration, and temperature are all considered. The adsorption kinetics results are adjusted to best fit the pseudo-second-order model. Langmuir-Freundlich and Langmuir models well describe the experimental data with an adsorption capacity of 472 mg g−1. The calculated thermodynamic data demonstrates that adsorption is spontaneous and endothermic. Desorption studies with methanol indicate that the adsorbent could successfully retain MV, even after five cycles. The statistical physics theory indicates the non-parallel orientation of the MV molecule’s adsorption. The adsorption energies varied from 13.99 to 17.60 kJ mol−1, revealing the physical adsorption systems. From these results, it can be considered that the raw Algerian bentonite sample tested herein is effective in removing MV from aqueous solutions and may be used as an alternative to high-cost commercial adsorbents.

Strategies for detecting land-use change on the River Tea SCI ecological corridor via satellite images

Land-cover change is often accompanied by land-use change, altering ecosystem services, species habitats and contributing to climate change. The general goal of this study is to analyse different sources of information to quantify land use and land cover change (LULCC) in the River Tea SCI (Galicia, NW Spain) in 2015–2023. The study area has multiple coverages with very low variability between them and with a great deal of fragmentation in the territory, which makes it difficult to obtain high accuracy levels. Land cover was classified using object-based image analysis (OBIA) Classification and Artificial Neural Network (ANN) methodologies based on images from the Sentinel-2 and Planet Labs (RapidEye and PlanetScope) multispectral satellite platforms. In addition, simulations for 2031 were carried out using different techniques based on post-classification. The highest accuracy obtained in this study is 80 %, for the Planet Labs data and the OBIA methodology. Using the same methodology and Sentinel-2 data the best accuracy obtained is around 70 %, with the Planet Labs data and the ANN developed the accuracy is around 55 %. Thus, the data used and the methodology followed influence the accuracy levels obtained in the classifications. The choice of what data and methodology are to be used depends on the goals pursued and on the characteristics of the study area. Finally, it is concluded that the geospatial data available are very useful for detecting and quantifying changes in land cover. It is confirmed that the methodology used contributes to territorial planning and sustainable forest management, facilitating future decisions and action plans in the governance of the region.

Optimizing Cold Metal Transfer-Wire Arc Additive Manufacturing Parameters for Enhanced Mechanical Properties and Microstructure of ER5356 Aluminum Alloy Using Artificial Neural Network and Response Surface Methodology

Optimizing Cold Metal Transfer-Wire Arc Additive Manufacturing Parameters for Enhanced Mechanical Properties and Microstructure of ER5356 Aluminum Alloy Using Artificial Neural Network and Response Surface Methodology

Analyzing Data Using Artificial Intelligence for Optimization of Electrospun Polyvinyl Alcohol-Averrhoa bilimbi Nanofiber Diameters

Our reserach described in this paper aimed to predict the diameter of PVA-Averrhoa bilimbi nanofibers by applying artificial intelligence (Artificial Neural Network and Response Surface Methodology) modeling techniques. Three independent variables, namely voltage, distance, and concentration of Averrhoa bilimbi extract, were considered while the dependent variable was the nanofiber diameter. Response Surface Methodology (RSM) yielded an R-squared value of 63.76%, whereas the Artificial Neural Network (ANN) demonstrated better predictive capability with an R-squared value of 99.36%. This study, we believe, presents a novel approach for predicting the diameter of PVA-Averrhoa bilimbi nanofibers using RSM and ANNs, to our knowledge, for the first time. Another novelty of this research was the use of artificial intelligence to calculate the effect of electrospinning treatments on nanofiber diameter. The model presented in this study is expected to be useful for practitioners, researchers, and technicians seeking to optimize nanofiber production parameters.

Analyzing tube arrangements of a finned-tube heat exchanger to optimize overall efficiency using artificial neural network and response surface methodology

Temperature is a critical factor in numerous equipment, industries, and daily life applications. Heat exchangers are essential devices that help regulate and optimize this factor. The finned-tube heat exchanger (FTHE) is widely favored due to its high efficiency in facilitating heat transfer between liquids and gases. Improving the performance of FTHE can significantly provide thermal requirements in industrial and engineering processes and reduce energy consumption. The present research includes a numerical study of an FTHE and the optimization of the performances based on the input variables by response surface methodology (RSM) and artificial neural network (ANN) methods. The tubes' transverse and longitudinal pitches and inlet Reynolds number were selected as input variables. Also, the examined responses were the Colburn and friction factors. Changing tubes' pitches makes it possible to significantly affect the thermo-hydraulic performance without incurring additional costs and processes. The acquired results showed that the responses predicted by the models are very close to the numerical results, which indicates the high accuracy and validity of these models. According to the results, the optimum heat exchanger's efficiency index was obtained at Pt = 26.128 mm, Pl = 28 mm, and Re = 800. It was also observed that the overall performance of the optimal design is 185 % higher than the weakest FTHE.

Improving prediction accuracy of open shop scheduling problems using hybrid artificial neural network and genetic algorithm

Scheduling issues are typically classified as constrained optimization problems that examine the allocation of machines and the sequence in which tasks are processed. Regarding the existence of one machine, identification of works processing sequence forms a complete time schedule. Therefore, following a review of previous works, the goal of the present study is designing a mathematical model for open shop scheduling (OSS) problems using different machines aiming at minimizing the maximum time required to complete the works using an artificial neural network (ANN) and genetic algorithm (GA). The research data were driven from a Shoe company carried out between the years 2019 and 2020. The GA and ANN methodologies were employed to analyze and forecast the scheduling of activities within the shoe manufacturing sector. The findings indicated that the probability associated with the third population of the GA was 0.15. Furthermore, an examination of the average values of standard error revealed that the neural network model outperformed in terms of predictive accuracy. The estimated minimum time necessary for task completion, as determined by the neural network, was calculated to be 0.96699, facilitating an optimal condition for meeting the established objectives.

Microplastic predictive modelling with the integration of Artificial Neural Networks and Hidden Markov Models (ANN-HMM).

Microplastic pollution poses a significant threat to our environment, necessitating effective predictive modelling approaches for better management and mitigation. In this study, we introduce a pioneering methodology that fuses the power of Artificial Neural Networks (ANN) and Hidden Markov Models (HMM) for microplastic predictive modelling. Leveraging a comprehensive dataset, our integrated model exhibits exceptional performance, with an Accuracy of 0.96, Precision of 0.96, Recall of 0.97, and an F1 Score of 0.96. The achieved Accuracy underscores the model's proficiency in distinguishing microplastic and non-microplastic entities, promising robust and reliable predictions. Precision, as a measure of correct positive identifications, demonstrates our model's effectiveness in minimizing false positives, a crucial aspect for environmental monitoring. Moreover, the perfect Recall score signifies the model's ability to detect all relevant microplastic instances, addressing concerns about false negatives. The F1 Score encapsulates this dual proficiency, showcasing a harmonious trade-off between precision and recall. Our research not only advances the field of microplastic prediction but also highlights the potential of synergizing ANN and HMM methodologies for comprehensive environmental assessments. The reported performance metrics underscore the practical applicability of our approach, offering a valuable tool for tackling the pervasive issue of microplastic pollution and fostering proactive environmental stewardship.

Development of Soft Computing-based Predictive Tools for Estimating the Young Modulus of Weak Rocks

The deformation characteristics of rocks are of vital importance in addressing most geomechanical issues as they are one of the most critical input parameters in rock engineering analyses. For this reason, robust forecasting models are required when analysing the stability of tunnels, slopes, mine galleries, and other underground excavations. In this research, novel predictive models are proposed to estimate the tangential Young modulus (Eti) of weak rocks. To achieve this, an extensive literature review is performed to obtain a comprehensive database including critical physico-mechanical properties of various weak rocks. Thanks to the advantages of soft computing methods such as genetic algorithm (GA), adaptive neuro-fuzzy inference system (ANFIS), artificial neural networks (ANN) and multivariate adaptive regression splines (MARS), novel predictive models are established. The effectiveness of the developed predictive models is investigated using various statistical measures and it is concluded that empirical models utilizing ANN and ANFIS methodologies are the most effective tools for estimating the Eti of weak rocks. In addition, a practical design chart is also developed for assessing the Eti of weak rocks.

Effect of operating variables on isopropanol absorbed by triethylene glycol solution and discussions on the mass transfer coefficient predicted by artificial neural network

The mass transfer performance of isopropanol (IPA) absorption by triethylene glycol (TEG) aqueous solution in a spray-bed absorber was investigated, and the mass transfer coefficients predicted by an artificial neural network (ANN) and response surface methodology were compared in this study. The operating variables affecting the mass transfer performance of a spray-bed absorber include the gas and TEG fluxes, concentration of IPA, concentration of TEG aqueous solution, and operating temperature. The experimental results demonstrated that the mass transfer coefficient of a spray-bed absorber increased with increases in gas and liquid fluxes and decreased with increases in concentration of IPA. The Back-Propagation algorithm in the ANN model was selected to predict the mass transfer coefficients of the spray–bed absorption system. Hidden layers with 2, 4 and 6 neurons respectively were set to analyze the effect of the number of neurons on the target value, and then different transfer functions were used to compare the accuracy in predicting mass transfer coefficient. Finally, the influences of the number of hidden layers and different learning rules on the prediction of mass transfer coefficients were examined. The simulated results showed that the average error of the ANN using the TanhAxon transfer function, the Conjugate Gradient learning rule, two hidden layers, and 6 neurons in each hidden layer was within 1%. The results not only verified the successful establishment of the experimental device but also proved the accurate prediction of mass transfer coefficients by the ANN model.

Optimizing Wood Composite Drilling with Artificial Neural Network and Response Surface Methodology

Optimizing Wood Composite Drilling with Artificial Neural Network and Response Surface Methodology

4D printing and optimization of biocompatible poly lactic acid/poly methyl methacrylate blends for enhanced shape memory and mechanical properties

This study introduces a novel approach to 4D printing of biocompatible Poly lactic acid (PLA)/poly methyl methacrylate (PMMA) blends using Artificial Neural Network (ANN) and Response Surface Methodology (RSM). The goal is to optimize PMMA content, nozzle temperature, raster angle, and printing speed to enhance shape memory properties and mechanical strength. The materials, PLA and PMMA, are melt-blended and 4D printed using a pellet-based 3D printer. Differential Scanning Calorimetry (DSC) and Dynamic Mechanical Thermal Analysis (DMTA) assess the thermal behavior and compatibility of the blends. The ANN model demonstrates superior prediction accuracy and generalization capability compared to the RSM model. Experimental results show a shape recovery ratio of 100% and an ultimate tensile strength of 65.2 MPa, significantly higher than pure PLA. A bio-screw, 4D printed with optimized parameters, demonstrates excellent mechanical properties and shape memory behavior, suitable for biomedical applications such as orthopaedics and dental implants. This research presents an innovative method for 4D printing PLA/PMMA blends, highlighting their potential in creating advanced, high-performance biocompatible materials for medical use.

Catalytic converter performance prediction and engine optimization when powered by diisopropyl ether/gasoline blends: Combined application of response surface methodology and artificial neural network

The main objective of this study is to develop a neural network model to predict high-accuracy responses and optimize the input parameters using response surface methodology to improve catalytic converter performance when powered by diisopropyl ether/gasoline blends. The engine exhaust gas was treated by a commercial catalytic converter by varying the brake power, engine speed, and compression ratio, and the pollutant levels were measured. Again, the experiment was repeated, and the exhaust gas was treated by a sucrose/alumina catalyst-coated converter and the results were compared. Experimentally obtained data were employed to develop a neural network and response surface methodology model. The developed artificial neural network yielded higher R2 values of over 0.9977 for commercial catalytic converter and over 0.99633 for sucrose/alumina catalyst-coated converter. Furthermore, mean square error values were less than 3 % for all the responses which indicates higher prediction accuracy. Similarly, the developed response surface methodology model exhibited a higher F-value and lower p-value which indicates the model is accurate. Moreover, the desirability factor of over 0.99 for both cases shows the model's high stability. The predicted optimum brake power of 8.01 kW at 2500 rpm and 9.5 compression ratio produced lesser emissions and the validation test error was found to be less than 4 %. Thus, the authors conclude that the combined application of artificial neural network and response surface methodology will be efficient in exhaust gas pollution level prediction and optimization.

Evaluation of Machine Learning and Traditional Methods for Estimating Compressive Strength of UHPC

This research provides a comparative analysis of the optimization of ultra-high-performance concrete (UHPC) using artificial neural network (ANN) and response surface methodology (RSM). By using ANN and RSM, the yield of UHPC was modeled and optimized as a function of 22 independent variables, including cement content, cement compressive strength, cement type, cement strength class, fly-ash, slag, silica-fume, nano-silica, limestone powder, sand, coarse aggregates, maximum aggregate size, quartz powder, water, super-plasticizers, polystyrene fiber, polystyrene fiber diameter, polystyrene fiber length, steel fiber content, steel fiber diameter, steel fiber length, and curing time. Two statistical parameters were examined based on their modeling, i.e., determination coefficient (R2) and mean square error (MSE). ANN and RSM were evaluated for their predictive and generalization capabilities using a different dataset from previously published research. Results show that RSM is computationally efficient and easy to interpret, whereas ANN is more accurate at predicting UHPC characteristics due to its nonlinear interactions. Results show that the ANN model (R = 0.95 and R2 = 0.91) and RSM model (R = 0.94, and R2 = 0.90) can predict UHPC compressive strength. The prediction error for optimal yield using an ANN and RSM was 3.5% and 7%, respectively. According to the ANN model’s sensitivity analysis, cement and water have a significant impact on compressive strength.

Accurate Prediction of Compression Index of Normally Consolidated Soils Using Artificial Neural Networks

The compression index (Cc) serves as a crucial parameter in predicting consolidation settlement in fine-grained soils, representing the slope of the void ratio logarithmic effective stress curve obtained from oedometer tests. However, traditional consolidation testing methods are notably time-consuming, typically spanning a 15-day period for preparation, execution, and parameter calculation, leading to significant delays in civil engineering projects. Therefore, there is an urgent need for effective methodologies to determine consolidation parameters within a shorter timeframe. Although various empirical formulas have been proposed over the years to correlate compressibility with soil parameters, none have reliably predicted the Cc across different datasets. In this study, to overcome this challenge, an alternative approach using artificial neural network (ANN) methodology to predict the compression index of fine-grained soils based on index properties is proposed. For this purpose, an ANN was trained and validated using a dataset consisting of 560 high and low- plasticity soil samples obtained from construction sites in various regions of Turkey over the last forty years, as well as soil borings in Istanbul. The modeling of artificial neural networks was performed using the Regression Learner program, which integrates with the Matlab 2023a software package and offers a user-friendly graphical interface for AI model development without coding. The data set, which was structured as a matrix with dimensions of 458 × 6, included input parameters such as the natural water content, liquid limit, plastic limit, plastic index and initial void ratio, as well as information on the compression index, which was the output variable. The developed ANN model showed an outstanding predictive performance when predicting the output of the test data, achieving an outstanding R2 score of 0.81. This underlines the potential of ANN methodologies to efficiently extract important data with fewer experiments and in less time, and offers promising applications in the field of geotechnical engineering.

Novel glassbox based explainable boosting machine for fault detection in electrical power transmission system.

The reliable operation of electrical power transmission systems is crucial for ensuring consumer's stable and uninterrupted electricity supply. Faults in electrical power transmission systems can lead to significant disruptions, economic losses, and potential safety hazards. A protective approach is essential for transmission lines to guard against faults caused by natural disturbances, short circuits, and open circuit issues. This study employs an advanced artificial neural network methodology for fault detection and classification, specifically distinguishing between single-phase fault and fault between all three phases and three-phase symmetrical fault. For fault data creation and analysis, we utilized a collection of line currents and voltages for different fault conditions, modelled in the MATLAB environment. Different fault scenarios with varied parameters are simulated to assess the applied method's detection ability. We analyzed the signal data time series analysis based on phase line current and phase line voltage. We employed SMOTE-based data oversampling to balance the dataset. Subsequently, we developed four advanced machine-learning models and one deep-learning model using signal data from line currents and voltage faults. We have proposed an optimized novel glassbox Explainable Boosting (EB) approach for fault detection. The proposed EB method incorporates the strengths of boosting and interpretable tree models. Simulation results affirm the high-efficiency scores of 99% in detecting and categorizing faults on transmission lines compared to traditional fault detection state-of-the-art methods. We conducted hyperparameter optimization and k-fold validations to enhance fault detection performance and validate our approach. We evaluated the computational complexity of fault detection models and augmented it with eXplainable Artificial Intelligence (XAI) analysis to illuminate the decision-making process of the proposed model for fault detection. Our proposed research presents a scalable and adaptable method for advancing smart grid technology, paving the way for more secure and efficient electrical power transmission systems.