A Global Modeling Pruning Ensemble Stacking With Deep Learning and Neural Network Meta-Learner for Passenger Train Delay Prediction

Streamflow forecasting is crucial in hydrology and hydraulic engineering since it is capable of optimizing water resource systems or planning future expansion. This study investigated the performances of three different soft computing methods, multilayer perceptron neural network (MLPNN), optimally pruned extreme learning machine (OP-ELM), and evolutionary polynomial regression (EPR) in forecasting daily streamflow. Data from three different stations, Soleyman Tange, Perorich Abad, and Ali Abad located on the Tajan River of Iran were used to estimate the daily streamflow. MLPNN model was employed to determine the optimal input combinations of each station implementing evaluation criteria. In both training and testing stages in the three stations, the results of comparison indicated that the EPR technique would generally perform more efficiently than MLPNN and OP-ELM models. EPR model represented the best performance to simulate the peak flow compared to MLPNN and OP-ELM models while the MLPNN provided significantly under/overestimations. EPR models which include explicit mathematical formulations are recommended for daily streamflow forecasting which is necessary in watershed hydrology management.

Experimental and numerical investigation of heat transfer and flow of water-based graphene oxide nanofluid in a double pipe heat exchanger using different artificial neural network models

The objective of this study was to develop a multilayer perceptron neural network (MLPNN) and radial basic function neural network (RBFNN) model to predict the dissolved oxygen (DO) at some selected locations at the Kainji hydropower reservoir, Nigeria. The neural networks (NN) model was developed using water quality data collected over a six-year period (2010 to 2015). The NN structure was designed and trained using the SPSS neural network toolbox. The input variables to the NN were: pH, temperature, chloride (Cl-), PO43-, NO3-, Fe2+, and electrical conductivity (EC), while the output was the DO. The performance evaluation of the model was carried out using the coefficient of correlation (r), mean square error (MSE) and mean relative error (MRE). A positive correlation was observed between the actual and simulated DO at the four locations. The results of the simulation showed that the application of the NN and multiple regression analysis to predict DO concentration in water gave satisfactory results for all the selected locations using the two NN modeling approaches. Thus it has been demonstrated that NN modeling tools and multiple regression analysis are very efficient and useful for the computation of water quality parameters.

Kinetic equilibrium reconstructions make use of profile information such as particle density and temperature measurements in addition to magnetics data to compute a self-consistent equilibrium. They are used in a multitude of physics-based modeling. This work develops a multi-layer perceptron (MLP) neural network (NN) model as a surrogate for kinetic Equilibrium Fitting (EFITs) and trains on the 2019 DIII-D discharge campaign database of kinetic equilibrium reconstructions. We investigate the impact of including various diagnostic data and machine actuator controls as input into the NN. When giving various categories of data as input into NN models that have been trained using those same categories of data, the predictions on multiple equilibrium reconstruction solutions (poloidal magnetic flux, global scalars, pressure profile, current profile) are highly accurate. When comparing different models with different diagnostics as input, the magnetics-only model outputs accurate kinetic profiles and the inclusion of additional data does not significantly impact the accuracy. When the NN is tasked with inferring only a single target such as the EFIT pressure profile or EFIT current profile, we see a large increase in the accuracy of the prediction of the kinetic profiles as more data is included. These results indicate that certain MLP NN configurations can be reasonably robust to different burning-plasma-relevant diagnostics depending on the accuracy requirements for equilibrium reconstruction tasks.

<div class="WordSection1"><p>Assessing the performance of an educational institute is a prime concern in an educational scenario. Educational Data Mining (EDM) considers several tasks originated from an educational context. One of the tasks identified is providing feedback for supporting instructors, administrators, teachers, course authors in decision making and thereby enable them to take appropriate remedial action. In this research, we have developed a prototype Neural Network Model which is trained to predict the performance of an educational institution. A Multilayer Perceptron Neural Network (MLP) model had been developed for this proposed research. The network is trained by back propagation algorithm. Data was obtained from a well-defined questionnaire consisting of 14 questions in the domains namely Academic Schedule, International Exposure, Jobs and Internship, Quality of the college, and Life at Campus. The results of these questions have been taken as inputs and performance of the institute has been considered as the output. To, validate the results generated by the network, statistical techniques have been used for the purpose. In this proposed research performance of an educational institution has been predicted. The results generated by the Neural Network and the statistical techniques have been compared in this research and it is observed that, both the methods have generated accurate results. The results have been considered based on the Normalized System Error (NSE) values of the network. A prototype Neural Network model has been developed to assess the performance of an educational institution.</p><strong></strong></div>

SUMMARY & CONCLUSIONSThe objective of this research is to assess the reliability characteristics of the Allison T-56 aircraft turbine system which is subjected to local GCC environment where sand erosion is a major factor in reducing turbine life. To utilize this study, mor than thirty years of local fielded data were used for failure rate prediction and validation. The work utilizes Artificial Neural Networks (ANN) employing feed-forward Back-Propagation (BP) and Radial Basis Functions (RBF) neural network algorithm to predict the general T-56 engine turbine failure rate and the failures which required an overhaul maintenance, The results show that the failure times predicted by the RBF model. is closer in agreement with actual failure date than BP model. By the end of the study, we forecast the general failure rate of T-56 Engine Turbine, the failures which required overhaul maintenance and six categorical failures using Multilayer Perceptron neural network (MLP) model on DTREG commercial software. The results also give an insight into the reliability of the engine turbine under actual operating conditions, which can be used by aircraft operators for assessing system and component failures and customizing the maintenance programs recommended by the manufacturer.

Experimental analysis and low-damage machining strategy for composite ultrasonic vibration-assisted grinding of silicon carbide based on DA-MLP-NSGA-II algorithm

Forecasting surface water temperature in lakes: A comparison of approaches

Research into the problem of classification of radar returns from the ionosphere has been taken up as a challenging task for the neural networks (NNs). It appears from the literature review that for the Multi layer Perceptron (MLP) NN trained with backpropagation, reported average classification accuracy was about 96% on the test instances. This paper investigates and designs an optimal classifier using a radial basis function (RBF) NN. Authors compare the performance of two NN configurations, namely a well-known MLP NN model and the proposed RBF NN model on the radar dataset collected from the published studies. It is shown that the proposed RBF NN, consistently, has 100% accuracy on bad instances and 99.1935% accuracy on good instances. The results show that the proposed RBF NN classifier clearly outperforms the MLP NN one in various performance measures such as MSE, NMSE, correlation coefficients, area under the ROC curve and classification accuracy on the testing datasets even after attempting different data partitions.

In the present study, several hybrids and coupled forms of machine learning algorithms were developed to provide accurate elastic modulus of recycled aggregate concrete’s (ERAC) estimation, called multilayer perceptron neural networks (MLPNN). For this gain, a comprehensive dataset was collected from literature containing 400 rows of samples. Networks with different hidden layer numbers were used to develop MLPNN (from one layer to three layers). The accuracy of MLPNN deeply related to neuron numbers of each hidden layer, wherein the present paper, three optimization algorithms; the arithmetic optimization algorithm (AOA), equilibrium optimizer (EO), and flow direction algorithm (FDA); are merged with MLPNN to obtain the ideal value of neuron numbers. Considering all models, all developed models have the acceptable ability for predicting the ERAC by the coefficient of determination of at least 0.9306 for the learning stage and 0.9411 for the examining stage. All in all, the MLPNN model with two numbers of hidden layers with a structure of 17-14-1 optimized with AOA can be proposed as the most appropriate model. The created MLPNN models may optimize material selection and quality control in building processes, reducing dependence on natural aggregates.

In Vitro Fertilization (IVF) is one of the scientifically known methods of infertility treatment. This study aimed at improving the performance of predicting the success of IVF using machine learning and its optimization through evolutionary algorithms. The Multilayer Perceptron Neural Network (MLP) were proposed to classify the infertility dataset. The Genetic algorithm was used to improve the performance of the Multilayer Perceptron Neural Network model. The proposed model was applied to a dataset including 594 eggs from 94 patients undergoing IVF, of which 318 were of good quality embryos and 276 were of lower quality embryos. For performance evaluation of the MLP model, an ROC curve analysis was conducted, and 10-fold cross-validation performed. The results revealed that this intelligent model has high efficiency with an accuracy of 96% for Multi-layer Perceptron neural network, which is promising compared to counterparts methods.

The article examines the dependence between the Russian ruble exchange rate and oil prices with the use of neural network modeling. The relevance of the study can be confirmed by the interest of the monetary authorities in modeling the dynamics of the exchange rate for developing monetary policy measures. The research objective of the article is the estimation of the relationship between the Russian ruble exchange rate and oil prices using multilayer perceptron and recurrent neural network models. Moreover, the influence of additional factors, including foreign exchange interventions and geopolitical risks, is estimated. The results show that neural networks provide sufficient accuracy in estimation of the target variable. Furthermore, during the periods with foreign exchange interventions and high geopolitical instability there was confirmed a decoupling of the examined variables. The modeled time series preserve non-linear nature of exchange rate data generating process, as well as the asymmetry in the reaction of the ruble exchange rate to oil price shocks. The hyperparameters selection, use of bootstrap and ensembles of neural networks provide more robust estimates and confidence intervals for the oil price elasticity of the ruble exchange rate. Therefore, the combination of the aforementioned methods makes it possible to draw meaningful economic conclusions based on the trained neural networks, avoiding the problem of neural network weights non-interpretability.

Iraq’s arid climate faces intricate challenges due to volatile precipitation patterns and limited water resources. Prediction of precipitation environmental isotopes using neural network techniques represents optimistic way in water science. This technique has demonstrated effectiveness in interpreting intricate data and producing significant insights on the water cycle. The present study aimed to develop the neural network model by using daily precipitation samples gathered from different sites across Iraq from 2010 to 2024. The input variables used for the ANN models include measurements of stable isotope in precipitation, specifically δ¹8O, δ2H and deuterium excess (d-excess). The Multi-Layer Perceptron (MLP) and Radial Basis Neural Network (RBNN) models are employed for comparison with observed values to assess model accuracy and robustness. Additionally, The performance of the applied models is evaluated using statistical metrics such as the coefficient of determination (R2), Mean Absolute Error (MAE), and Root Mean Square Error (RMSE) statistics. The results of the study demonstrate that the (RBNN) model outperforms the (MLP) in predictive accuracy and overall performance. Reduced error metrics (RMSE and MAE) and stronger explanatory power (R2) demonstrate the higher accuracy of RBNN, particularly when modelling data with complex spatial and temporal patterns. This demonstrates that, particularly in dynamic climates and environmental systems.

This study presents a methodology to improve monitoring of gas turbines (GTs) while not having the specific engine characteristics available by applying a nonparametric statistical modeling based on operational data. The goal is to quantify the relationship between the operational conditions and the GT performance parameters during normal operation. In this study, the Multi-Layer Perceptron (MLP) Neural Network (NN) model is used to develop so called baseline models of the GTs. To verify the generic applicability, two different GTs and several units for each GT type are evaluated. A methodology to both, selecting the appropriate models with regard to input and output parameters as well as validating the selection of the parameters is reported. The key result can be summarized as: MLPs can be used for statistical modeling of GTs with very high accuracy, which permits an accurate prediction of the performance parameters during different operational conditions. It is also shown that under certain conditions, a developed NN model can be transferred to a similar GT. In summary, this study shows that NN can be used to improve on-line monitoring of a GT, making it possible to detect trends in the measured data indicating a change in the health status of the engine, requiring operational data only for model development.

Complexity Analysis of Multilayer Perceptron Neural Network Embedded into a Wireless Sensor Network