Back Propagation Neural Network Method Research Articles

Using Backpropagation Neural Network for Polyvinylchloride Ceiling Price Modeling

Sales predictions on building material products today have applied an artificial neural network approach. One of the products of building material that need to be predicted for sales is polyvinylchloride (PVC) ceilings. Most companies haven’t implementing prediction technique for the sale of PVC ceilings, so this study aims to predict PVC ceiling sales with the backpropagation neural network (BPNN) method using the R algorithm. Unit gradients are calculated using the average absolute per cent error value (MAPE) to minimize the total square errors of network output. The results showed that the network architecture used was 4 to 6-1 and obtained an accuracy of 88% based on the lowest MAPE value.

Prediction of Student Graduation Rates using the Artificial Neural Network Backpropagation Method

This student graduation rate research focuses on analyzing academic performance with the main aim of identifying and distinguishing between students who graduate on time and those who graduate late. The application of data mining techniques in this research uses the neural network method, which is expected to offer deeper insight into the factors that influence students' graduation times. In this study, the neural network method was used to classify graduation data from 150 students. The results of this analysis were very encouraging, with 149 students identified as graduating on time and one student graduating late. The level of accuracy achieved in this classification is 98%, which shows the effectiveness of the neural network method in processing and analyzing academic data. These results confirm that neural networks are a powerful and reliable tool for predictive tasks like this. The successful use of neural networks in this study also proves their potential in broader educational applications, particularly in optimizing educational and intervention strategies. By understanding the characteristics of students who graduate on time versus those who graduate late, educators and administrators can design more effective programs to support student success. This is important not only to improve graduation statistics, but also to improve the overall educational experience for students.

Global Navigation Satellite System Receiver Positioning in Harsh Environments via Clock Bias Prediction by Empirical Mode Decomposition and Back Propagation Neural Network Method.

This paper proposes a novel method to improve the clock bias short-term prediction accuracy of navigation receivers then solve the problem of low positioning accuracy when the satellite signal quality deteriorates. Considering that the clock bias of a navigation receiver is equivalent to a virtual satellite, the predicted value of clock bias is used to assist navigation receivers in positioning. Consequently, a combined prediction method for navigation receiver clock bias based on Empirical Mode Decomposition (EMD) and Back Propagation Neural Network (BPNN) analysis theory is demonstrated. In view of systematic errors and random errors in the clock bias data from navigation receivers, the EMD method is used to decompose the clock bias data; then, the BPNN prediction method is used to establish a high-precision clock bias prediction model; finally, based on the clock bias prediction value, the three-dimensional positioning of the navigation receiver is realized by expanding the observation equation. The experimental results show that the proposed model is suitable for clock bias time series prediction and providing three-dimensional positioning information meets the requirements of navigation application in the harsh environment of only three satellites.

Projection of PT Aneka Tambang Tbk Share Risk Value Based on Backpropagation Artificial Neural Network Forecasting Result

PT Aneka Tambang Tbk (ANTAM) received an award as the most sought-after stock issuer in Indonesia in 2016. That stock continued to attract investors in 2022 due to a 105% increase in net profit and a 19% increase in sales from the previous year. Despite the upward trend, investors still had doubts due to the fluctuating movement of ANTAM's stock prices. Therefore, forecasting was needed to determine the future movement of stock prices. The Backpropagation Neural Network method had good capabilities for fluctuating data types. However, this method has the disadvantage of a lengthy iteration process. To handle this limitation, The Nguyen-Widrow weighted setting was applied to address this constraint. The expected Shortfall (ES) method used the forecasting results to measure investment risk. This research uses ANTAM stock closing price data from May 2, 2018, to May 31, 2023. Based on the analysis results, the best architecture was obtained with a configuration of 5-11-1, using Nguyen-Widrow weight initialization and a combination of a learning rate of 0.5 and momentum of 0.9. This architecture yielded a prediction error based on the Mean Absolute Percentage Error (MAPE) of 1.9947%. Risk measurement with the ES method based on the prediction for the next 60 periods showed that at a 95% confidence level, the risk value was 0.002181; at a 90% confidence level, it was 0.002165; at an 85% confidence level, it was 0.002148, and at an 80% confidence level, it was 0.002132.

The Prediction of Pervious Concrete Compressive Strength Based on a Convolutional Neural Network

To overcome limitations inherent in existing mechanical performance prediction models for pervious concrete, including material constraints, limited applicability, and inadequate accuracy, this study employs a deep learning approach to construct a Convolutional Neural Network (CNN) model with three convolutional modules. The primary objective of the model is to precisely predict the 28-day compressive strength of pervious concrete. Eight input variables, encompassing coarse and fine aggregate content, water content, admixture content, cement content, fly ash content, and silica fume content, were selected for the model. The dataset utilized for both model training and testing consists of 111 sample sets. To ensure the model’s coverage within the practical range of pervious concrete strength and to enhance its robustness in real-world applications, an additional 12 sets of experimental data were incorporated for training and testing. The research findings indicate that, in comparison to the conventional machine learning method of Backpropagation (BP) neural networks, the developed CNN prediction model in this paper demonstrates a higher coefficient of determination, reaching 0.938, on the test dataset. The mean absolute percentage error is 9.13%, signifying that the proposed prediction model exhibits notable accuracy and universality in predicting the 28-day compressive strength of pervious concrete, regardless of the materials used in its preparation.

A Production Prediction Model of Tight Gas Well Optimized with a Back Propagation (BP) Neural Network Based on the Sparrow Search Algorithm

The production of tight gas wells decreases rapidly, and the traditional method is difficult to accurately predict the production of tight gas wells. At present, intelligent algorithms based on big data have been applied in oil and gas well production prediction, but there are still some technical problems. For example, the traditional error back propagation neural network (BP) still has the problem of finding the local optimal value, resulting in low prediction accuracy. In order to solve this problem, this paper establishes the output prediction method of BP neural network optimized with the sparrow search algorithm (SSA), and optimizes the hyperparameters of BP network such as activation function, training function, hidden layer, and node number based on examples, and constructs a high-precision SSA-BP neural network model. Data from 20 tight gas wells, the SSA-BP neural network model, Hongyuan model, and Arps model are predicted and compared. The results indicate that when the proportion of the predicted data is 20%, the SSA-BP model predicts an average absolute mean percentage error of 20.16%. When the proportion of forecast data is 10% of the total data, the SSA-BP algorithm has high accuracy and high stability. When the proportion of predicted data is 10%, the mean absolute average percentage error is 3.97%, which provides a new method for tight gas well productivity prediction.

Post-cyclic behavior of Haikou marine clay and their predictions using back-propagation neural network model

The evaluation of post-cyclic behavior of soft clay is critical for the lifetime safety and performance assessment of various infrastructures built on/in it. In this study, by employing the marine clays collected from Haikou, China, a series of both conventional shear tests and post-cyclic shear tests under undrained condition were performed using triaxial setup to investigate the post-cyclic behavior of the clay, with the investigation foci placed on the post-cyclic shear stress and secant modulus. Some important factors, namely confining pressure, loading frequency, cyclic shear strain amplitude and post-cyclic shear strain, were taken into consideration. From both the results from the tests and the parameter sensitivity analysis using random forest method, the confining pressure and cyclic shear strain amplitude were found to be evidently influential on the post-cyclic behavior of the clay, while the influence of loading frequency was relatively insignificant. In addition, an intelligent model was established based on the back-propagation neural network method to predict the post-cyclic behavior of Haikou marine clay, which was proven to be more favorable as compared to other two prediction methods developed based on multi-variate regression analysis and random forest algorithm.

Underwater Sound Speed Field Forecasting Based on the Least Square Support Vector Machine

Underwater sound speed is one of the most significant factors that affects high-accuracy underwater acoustic positioning and navigation. Due to its complex temporal variation, the forecasting of the underwater sound speed field (SSF) becomes a challenging task. Taking advantage of machine learning methods, we propose a new method for SSF forecasting based on the least square support vector machine (LSSVM) and a multi-parameter model, aiming to enhance the forecasting accuracy of underwater SSF with hourly resolution. We first use a matching extension method to standardize profile data and train the LSSVM with the parameters of observation time, temperature, salinity, and depth. We then employ radial basis function kernels to construct the forecasting model of SSF. We validate the feasibility and effectiveness of the LSSVM model by comparing it with the polynomial fitting (PF) and back propagation neural network (BPNN) methods, using hourly data obtained from the measured data and open data. The results show that the means of the root mean square for the LSSVM based on the observation time parameter and the LSSVM based on the multi-parameter model achieve 0.51 m/s and 0.45 m/s, respectively, presenting a significant improvement compared with the PF (0.82 m/s) and BPNN (0.76 m/s) methods.

Data-driven approaches for predicting performance degradation of solid oxide fuel cells system considering prolonged operation and shutdown accumulation effect

Solid oxide fuel cell system is widely acknowledged as the leading alternative energy generation system in the field. Due to their high efficiency, low emissions, low noise, and various other advantages, solid oxide fuel cell systems are being considered for use in automobiles as a replacement for traditional internal combustion engines. However, prolonged operation and abnormal shutdowns can lead to performance degradation, which affects the efficiency, stability, and lifespan of stack. In the context of prolonged operation, considering the fluctuations in stack performance parameters and balance of plant, several regression models based on voltage parameters are established to accurately predict changes in stack performance. The results reveal that the genetic algorithm optimized backpropagation neural network model is highly sensitive in predicting the system performance degradation. Further analysis reveals that abnormal shutdowns can cause system performance fluctuations. As a result, the number and duration of shutdowns are incorporated into genetic algorithm optimized backpropagation neural network model for analysis. This model exhibits the best performance degradation evolution prediction compared to the long short-term memory and particle swarm optimization backpropagation neural network methods. This finding is important for monitoring and controlling the operational status of stack.

A decision-aid system for subway microenvironment health risk intervention based on backpropagation neural network and permutation feature importance method

Subway has become an important commute mode in daily life. However, subway microenvironment poses a threat to people's health and there is growing concern about its health risks. In this paper, a decision-aid system for subway microenvironment health risk intervention was proposed based on the combination of backpropagation neural network algorithm and permutation feature importance method. The non-linear relationship between indicators and the risks of subway microenvironment was established using the neural network. On this basis, the implementation of stepwise intervention, neural network prediction, and the generation of strategies was achieved utilizing MATLAB programming. Moreover, a subway station in Nanjing was taken as an example to illustrate the implementation and verify the effectiveness of the system. With the help of the system, health risks in the subway microenvironment can be systematically analyzed, efficiently optimized, and gradually solved. Our research contributes to breaking the deadlock in subway microenvironment intervention, which has long been confined to predicting and summarizing strategies. It provides more objective and effective intervention strategies for the health risk management of subway microenvironments.

Physics-informed data assimilation model for displacement prediction of hydrodynamic pressure-driven landslide

Due to the complex and non-linear characteristics of landslide evolution, the deformation law of landslides is always hard to be predicted. Considering the dynamical evolution process of landslide physical mechanism, a Physics-Informed Data Assimilation (PIDA) method is first proposed for landslide displacement forecasting. It can enhance the prediction ability of the physical forecasting model, and solves the problem of ignoring the physical significance of landslides in mathematical prediction model. Additionally, the superiority of Particle Filter (PF) in nonlinear and non-Gaussian system determines its application in complex problem. Aiming at the hydrodynamic pressure-driven landslide, the Sliding Zone Deterioration (SZD) model is constructed as the state equation of the PF algorithm. And the observation equation is formed by the relationship between landslide deformation and monitoring datasets. The physics-informed landslide displacement prediction model serves as a bridge, consisting of the shear strength parameter dynamic updated model and physical mechanism analysis model, for key parameters feedback, modification, and update. Comparing the results to traditional Long Short-Term Memory (LSTM) and Back Propagation Neural Network (BPNN) methods, the PIDA method combining triggering and instability mechanisms performs better in accuracy. Moreover, it enables dynamic and reliable landslide deformation prediction.

A hybrid reliability assessment method for slab track with changing structural parameters

Changes in the dynamic properties of high-speed rail (HSR) slab track structures can have a great impact on train safety and ride quality. However, the dynamic response of wheel–rail systems, which is related to operational safety, has rarely been considered in existing service reliability assessments of track structures. To solve this problem, a hybrid method was developed, with the serviceability limit state (SLS) first defined with respect to the derailment coefficient (DC) and wheel unloading rate (WUR). In calculating the reliability index, the response surface method and the first-order reliability method were used to solve the implicit expression of wheel–rail force in the SLS equation. To reduce the computation cost, a surrogate model expressing the non-linear mapping of the wheel–rail interaction is proposed. The computation efficiency and accuracy of the hybrid method were compared with Monte Carlo simulation and a back-propagation neural network (BPNN). The computation time of the hybrid method was only 1/8.4 of the BPNN method, while the accuracy of the reliability index was 98% for the DC and 97% for the WUR. The hybrid method was used to assess the reliability of a typical slab track structure under changing stiffness and damping coefficients of the fasteners, cement asphalt mortar and foundation. The stiffness and damping of the fasteners had the most impact on wheel–rail dynamics and track reliability. This research provides new insights into reliability assessments for HSR slab track with respect to train operational safety.

A comparative study on the difference of color space conversion based on table look-up method and neural network

A comparative study on the difference of color space conversion based on table look-up method and neural network

SOUL-Net: A Sparse and Low-Rank Unrolling Network for Spectral CT Image Reconstruction.

Spectral computed tomography (CT) is an emerging technology, that generates a multienergy attenuation map for the interior of an object and extends the traditional image volume into a 4-D form. Compared with traditional CT based on energy-integrating detectors, spectral CT can make full use of spectral information, resulting in high resolution and providing accurate material quantification. Numerous model-based iterative reconstruction methods have been proposed for spectral CT reconstruction. However, these methods usually suffer from difficulties such as laborious parameter selection and expensive computational costs. In addition, due to the image similarity of different energy bins, spectral CT usually implies a strong low-rank prior, which has been widely adopted in current iterative reconstruction models. Singular value thresholding (SVT) is an effective algorithm to solve the low-rank constrained model. However, the SVT method requires a manual selection of thresholds, which may lead to suboptimal results. To relieve these problems, in this article, we propose a sparse and low-rank unrolling network (SOUL-Net) for spectral CT image reconstruction, that learns the parameters and thresholds in a data-driven manner. Furthermore, a Taylor expansion-based neural network backpropagation method is introduced to improve the numerical stability. The qualitative and quantitative results demonstrate that the proposed method outperforms several representative state-of-the-art algorithms in terms of detail preservation and artifact reduction.

Improved back propagation neural network method for predicting sulfur content in hot metal

Blast furnace smelting is a traditional iron-making process. Its product, hot metal, is an important raw material for the production of steel. Steelmaking efficiency can be improved and steel product quality can be stabilized by using proper hot metal. Sulfur is an important indicator reflecting the quality of hot metal, it is necessary to establish an accurate prediction model to predict the sulfur content of hot metal, to effectively guide the production process. There is a non-linear relationship among the factors influencing the desulfurization effect during the blast furnace smelting process, and the back propagation neural network (BPNN) model has a strong ability to solve nonlinear problems. However, BPNN has the disadvantages of slow convergence speed and easy to fall into local minima. To improve the prediction accuracy, an improved algorithm combining Kmeans and BPNN is proposed in this paper. The study showed that compared with the BPNN model and case-based reasoning (CBR) model, the Kmeans-BPNN model has the lowest RMSE and MAPE values, which indicates a high degree of fit and a low degree of dispersion. The Kmeans-BPNN model has the largest HR value, which indicates the highest prediction accuracy. The proposed Kmeans-BPNN prediction model achieves a hit rate of 96%, which is 4.5% higher than before the improvement. It can effectively improve the prediction accuracy of hot metal sulfur content.

Gear walk instability caused by brake-disc friction characteristics

A multi-body dynamic rigid-flexible coupling model of landing gear is established to study the gear walk instability caused by the friction characteristics of the brake disc. After validating the model with the experimental results, the influence of the landing gear structure and braking system parameters on gear walk is further investigated. Among the above factors, the slope of the graph for the friction coefficient of the brake disc and the relative velocity of brake stators and rotors is the most influential factor on gear walk instability. Phase trajectory analysis verifies that gear walk occurs when the coupling of multiple factors causes the system to exhibit an equivalent negative damping trend. To consider a more realistic braking case, a back propagation neural network method is employed to describe the nonlinear behavior of the friction coefficient of the brake disc. With the realistic nonlinear model of the friction coefficient, the maximum error in predicting the braking torque is less than 10% and the effect of the brake disc temperature on gear walk is performed. The results reveal that a more negative friction slope may contribute to a more severe unstable gear walk, and reducing the braking pressure is an effective approach to avoid gear walk, which provides help for future braking system design.

Parameterization of optical properties for liquid cloud droplets containing black carbon based on neural network.

This paper introduces a novel back propagation (BP) neural network method to accurately characterize optical properties of liquid cloud droplets, including black carbon. The model establishes relationships between black carbon volume fraction, wavelength, cloud effective radius, and optical properties. Evaluated on a test set, the value of the root mean square error (RMSE) of the asymmetry factor, extinction coefficient, single-scattering albedo, and the first 4 moments of the Legendre expansion of the phase function are less than 0.003, with the maximum mean relative error (MRE) reaching 0.2%, which are all better than the traditional method that only uses polynomials to fit the relationship between the effective radius and optical properties. Notably, the BP neural network significantly compresses the optical property database size by 37,800 times. Radiative transfer simulations indicate that mixing black carbon particles in water clouds reduces the top-of-atmosphere (TOA) reflectance and heats the atmosphere. However, if the volume fraction of black carbon is less than 10-6, the black carbon mixed in the water cloud has a tiny effect on the simulated TOA reflectance.

Improving maize residue cover estimation with the combined use of optical and SAR remote sensing images

Conservation tillage is an important conservation measure for arable land in modern agricultural production, which plays an essential role in protecting black soil and improving the quality of arable land. The estimation of maize residue cover (MRC) can be used to obtain the spatial distribution characteristics of conservation tillage, which is essential for government departments to promote conservation tillage technology and understand the implementation of it. In this paper the southern part of the Songnen Plain was used as the study area, and Sentinel-2 MSI images and Sentinel-1 SAR images were used as data sources to correlate the spectral indices and radar backscatter coefficients with the field sampling data in the study area. The MRC estimation model of the study area was constructed using the Random Forest (RF) model, the Multiple Linear Stepwise Regression (MLSR) model, and Back Propagation Neural Network (BPNN) model, respectively. The results of the study showed that the correlation coefficients of normalized difference tillage index (NDTI), simple tillage index (STI), normalized difference index (NDI5), NDI7, shortwave infrared normalized difference residue index (SINDRI), normalized difference senescent vegetation index (NDSVI), normalized difference residue index 2 (NDRI2), NDRI3, NDRI4, NDRI5, NDRI6, NDRI7, NDRI8, NDRI9, and MRC in the study area were greater than 0.4, and the correlation coefficients were higher for NDTI and STI, which reached 0.861 and 0.860, respectively. The correlation coefficient between VV and MRC was 0.56 and between VH and MRC was 0.594. We used MLSR, RF, and BPNN methods in combination with Sentinel-2 MSI images and Sentinel-1 SAR images for MRC estimation. The synergistic use of Sentinel-2 MSI images and Sentinel-1 SAR images helped to improve the accuracy of the MRC estimation models and the correlation coefficient R2 of all three models to greater than 0.8. Based on the statistical analysis of remote sensing estimation results, we found that the average value of the MRC of the maize growing areas in Changchun, Siping, and eastern Songyuan in November 2020 was 66%, and 2% of farmland in the study area had a MRC of less than 30%.

Application of Artificial Neural Network in Predicting Direct Economic Losses Due to Earthquake

Accurately predicting the direct economic losses caused by earthquakes is important for policy makers for disaster budgets. Before a disaster strikes, it is important to consider the public policy costsassociated with disaster relief and recovery. The aim of this study is to provide a risk assessment approach, which can benefit all parties involved. Artificial neural networks are widely used for time seriesforecasting, especially financial forecasting. Therefore, this study proposes a cutting-edge forecastingmethod such as backpropagation neural network (BPNN) and other prediction methods: neural networkautoregressive (NNAR) and ARIMA-GARCH to obtain the best prediction results. This paper appliesinterpolation data to increase the amount of data used. Two interpolations were applied to amplify theoriginal small sample with virtual points, namely cubic splines and further piecewise interpolation using. The results of this study are the cubic spline interpolation is the most effective way to solve thesmall sampling problem to predict direct economic losses due to the Indonesian earthquake and theBPNN method outperforms other traditional methods with an RMSE of 0.024 in the training period and0.174 in the testing period, significantly lower than other methods. The results of this research can beused as reference material for the government in estimating the level of earthquake losses and can beused to develop risk reduction strategies

Machine learning–based optimal design of the in-situ leaching process parameter (ISLPP) for the acid in-situ leaching of uranium

The migration process of leached uranium in the in-situ leaching of uranium is considered a typical reactive transport problem. During this process, the lixiviant concentration and injection rate are important in-situ leaching process parameters (ISLPP) to efficiently recover uranium. However, several uncertain factors affect the outcomes of the ISLPP design. In addition, the repeated use of the reactive transport model (RTM) for investigating the acid in-situ leaching of uranium with the application of the Monte Carlo method leads to a substantial computational load. For this reason, a machine learning (ML)–based surrogate model was developed with the backpropagation neural network (BPNN) method to replace the RTM under the condition of uncertain parameters. Moreover, the simulated annealing optimisation model for ISLPP was created based on the proposed surrogate model. The optimal ISLPP was achieved that generated maximum profits from uranium recovery under different lixiviant prices, uranium prices and exploitation times. The optimal design framework of ISLPP based on the proposed ML algorithm was then applied in the Bayan-Uul sandstone-type uranium deposit in Inner Mongolia, China. From our analysis, it was demonstrated that the ML-based surrogate model exhibited great fitness with the RTM. The optimal results of the ISLPP indicated that the lixiviant concentration and injection rate could be adjusted based on the fluctuations in lixiviant price, uranium price and exploitation time. If the prices of sulphuric acid were high, a specific concentration of hydrogen peroxide could be injected into the injection well to promote the oxidation and dissolution of the uranium ore to increase the income from the uranium recovery. The optimisation model can also use the ISLPP scheme to boost the revenues from different lixiviant prices, uranium prices and exploitation times under the uncertainty of porosity, illustrating the applicability of the ML-based optimal design method of ISLPP in ISL mining. A general framework for developing surrogate models, as well as for conducting uncertainty analyses for a wide range of groundwater models was proposed here yielding valuable insights.