BP Neural Network Research Articles

An improved method for predicting the permeability of Chang 7 shale oil reservoirs based on lithology and hydraulic flow units

Accurate permeability characterization is essential for evaluating shale oil reservoirs quality. Chang 7 shale oil reservoirs, with diverse lithologies, low porosities and permeabilities, complex pore structures, and strong heterogeneity, pose challenges for traditional permeability prediction methods. This paper proposes an improved method that integrates lithology and hydraulic flow units (HFUs) with conventional logs for enhanced permeability prediction. The reservoir characteristics of the Chang 7 member in the Sai 392 area were analyzed, investigating relationships among lithology, porosity, pore structure parameters, and permeability. Lithology was identified by combining imaging logging and lithologic reconstruction curves (Ilith), with Ilith value boundaries of − 0.5, 1.5, and 3 for fine sandstone, argillaceous siltstone, silty mudstone, and mudstone, respectively. A back-propagation neural network predicted the flow zone index using lithological evaluation parameters and logging data, enabling HFU dentification and permeability model establishment. In this case study, the improved method was successfully implemented to accurately predict the permeability of the Chang 7 member shale oil reservoirs. Almost all data in the crossplot of the predicted versus measured permeability are within ± 0.17 uncertainty boundaries, with a root mean square error lower than 0.051. The results demonstrate that the improved method can effectively and accurately predict the permeability of the Chang 7 member shale oil reservoirs in the Sai 392 area.

Performance change prediction of stoner pipeline simulator simulation centrifugal compressor based on Whale Optimization Algorithm-enhanced back-propagation neural network

Performance change prediction of stoner pipeline simulator simulation centrifugal compressor based on Whale Optimization Algorithm-enhanced back-propagation neural network

Traffic conflict-based crash risk estimation: machine learning meets extreme value theory

This study proposed a hybrid framework integrating machine learning and extreme value theory (EVT) model for traffic conflict-based crash risk estimation. The threshold of conflict extremes was determined at first through peak over threshold plots. Secondly, the backpropagation neural network (BPNN) was employed to predict frequencies of conflict extremes of different severity levels based on dynamic traffic parameters, followed by three sampling methods to convert frequencies into extreme values. Lastly, the extreme values were fitted by generalized Pareto distribution (GPD) for crash risk estimation. The proposed framework was applied to a signalized intersection at the signal cycle level. The results show that 0.95s is an appropriate threshold to define conflict extremes. Both the BPNN prediction and histogram fitting of three sampling methods show high accuracy, and the estimated crash risks from the predicted conflict extremes are close to those estimated from observed conflict extremes. Moreover, the estimated crash risks are relatively robust to prediction errors in the BPNN model.

Improved Credit Card Fraud Detection Employing Probabilistic Back Propagation in Neural Networks

Improved Credit Card Fraud Detection Employing Probabilistic Back Propagation in Neural Networks

Interactive Brain Edema Tumor Segmentation In Different Modality Images With A Combined Novel Neural Network

MR/CT images are facing the difficulties during investigation of tumours in brain part especially in edema portions. Different modalities has lot of information to measure and find out the exact boundary range. MRI and CT images are playing a vital role in medical field to find out and curing the tumour in any area. If the image has big in size and has high amount of information of the MRI AND CT images, then it takes the time to register, filter and segment that images. This research article represents a programmed brain tumour detection technique to expand the exactness and yield and reduction the finding time. An essential role in the investigation of tumours is played by precise identification of the location and size of the brain tumour. The three stages of the analytic strategy are pre-handling of the MRI images, extraction, and grouping. The highlights are segregated based on Dual-Tree Complex wavelet modification following image histogram balance (DTCWT). At the very end, Back Propagation Neural Networks (BPN) is employed to explain the normal and dysfunctional brains. The Spatial Fuzzy K-Means Clustering effectively estimates tumour detection.

Leveraging Artificial Intelligence and Clinical Laboratory Evidence to Advance Mobile Health Applications in Ophthalmology: Taking the Ocular Surface Disease as a Case Study

ABSTRACTBackgroundThe advent of mobile health (mHealth) applications has fundamentally transformed the healthcare landscape, particularly within the field of ophthalmology, by providing unprecedented opportunities for remote diagnosis, monitoring, and treatment. Ocular surface diseases, including dry eye disease (DED), are the most common eye diseases that can be detected by mHealth applications. However, most remote artificial intelligence (AI) systems for ocular surface disease detection are predominantly based on self‐reported data collected through interviews, which lack the rigor of clinical evidence. These constraints underscore the need to develop robust, evidence‐based AI frameworks that incorporate objective health indicators to improve the reliability and clinical utility of remote health applications.MethodsTwo novel deep learning (DL) models, YoloTR and YoloMBTR, were developed to detect key ocular surface indicators (OSIs), including tear meniscus height (TMH), non‐invasive Keratograph break‐up time (NIKBUT), ocular redness, lipid layer, and trichiasis. Additionally, back propagation neural networks (BPNN) and universal network for image segmentation (U‐Net) were employed for image classification and segmentation of meibomian gland images to predict Demodex mite infections. These models were trained on a large dataset from high‐resolution devices, including Keratograph 5M and various mobile platforms (Huawei, Apple, and Xiaomi).ResultsThe proposed DL models of YoloMBTR and YoloTR outperformed baseline you only look once (YOLO) models (Yolov5n, Yolov6n, and Yolov8n) across multiple performance metrics, including test average precision (AP), validation AP, and overall accuracy. These two models also exhibit superior performance compared to machine plug‐in models in KG5M when benchmarked against the gold standard. Using Python's Matplotlib for visualization and SPSS for statistical analysis, this study introduces an innovative proof‐of‐concept framework leveraging quantitative AI analysis to address critical challenges in ophthalmology. By integrating advanced DL models, the framework offers a robust approach for detecting and quantifying OSIs with a high degree of precision. This methodological advancement bridges the gap between AI‐driven diagnostics and clinical ophthalmology by translating complex ocular data into actionable insights.ConclusionsIntegrating AI with clinical laboratory data holds significant potential for advancing mobile eye health (MeHealth), particularly in detecting OSIs. This study aims to explore this integration, focusing on improving diagnostic accuracy and accessibility. This study demonstrates the potential of AI‐driven tools in ophthalmic diagnostics, paving the way for reliable, evidence‐based solutions in remote patient monitoring and continuous care. The results contribute to the foundation of AI‐powered health systems that can extend beyond ophthalmology, improving healthcare accessibility and patient outcomes across various domains.

Remote Sensing Inversion Model of Water Quality Parameters Based on BP Neural Network and Spatial Distribution Analysis in the Middle Reaches of the Yangtze River Basin in China

Remote Sensing Inversion Model of Water Quality Parameters Based on BP Neural Network and Spatial Distribution Analysis in the Middle Reaches of the Yangtze River Basin in China

Enhanced Hybrid Feature Selection with HRFBP Algorithm for Medical Data Classification

Medical industry produces a significant portion of data whereas by adopting various machine learning models which can make accurate predictions about public healthcare that can be generalised. Machine learning is ideal for examining complex high dimensional data because of its flexibility and automation compared to conventional approaches. Still, feature extraction is not performed effectively and hence overall classifier accuracy is reduced considerably. To overcome the above mentioned problems, in this work, Hybrid Random Forest with Back Propagation (HRFBP) neural network algorithm is proposed. Initially, the datasets are collected which is preprocessed using K-Means Clustering (KMC) algorithm. It is used to handle the missing values and error rates efficiently. Then, the feature extraction is done by using Modified Principal Component Analysis (MPCA) which is focused to extract the significant features from the given medical dataset. After that, the feature selection is done by using EFS algorithm which generates best fitness values via objective function. EFS is done dependent on integrating numerous FS rather than a single FS to handle the FS issue. The possibility of EFS method is that merging the outcomes of a various single FS methods like Entropy Elephant Herding Optimization (EEHO), Adaptive Firefly Optimization Algorithm (AFOA) and Entropy Butterfly Optimization Algorithm (EBFO) acquire improved outcomes rather than utilizing a single FS methodology. Finally, the medical dataset classification is performed using HRFBP algorithm. The HRFBP algorithm performs training and testing process which learns a set of weights for prediction for the class label of features. HRFBP increases the classifier accuracy and reduces the error rates prominently. From the experimental result, it concluded that the proposed HRFBP algorithm provides better performance in terms of the higher accuracy, sensitivity, specificity and lower execution time rather than the existing algorithms

4D Track Prediction Based on BP Neural Network Optimized by Improved Sparrow Algorithm

The prediction accuracy of 4D (four-dimensional) trajectory is crucial for aviation safety and air traffic management. Firstly, the sine chaotic mapping is employed to enhance the sparrow search algorithm (Sine-SSA). This enhanced algorithm optimizes the threshold parameters of the BP (back propagation) neural network (Sine-SSA-BP), thereby improving the quality of the initial solution and enhancing global search capability. Secondly, the optimal weight thresholds obtained from the Sine-SSA algorithm are integrated into the BP neural network to boost its performance. Subsequently, the 4D trajectory data of the aircraft serve as input variables for the Sine-SSA-BP prediction model to conduct trajectory predictions. Finally, the prediction results from three models are compared against the actual aircraft trajectory. It is found that within the specified time series, the errors in longitude, latitude, and altitude for the Sine-SSA-BP prediction model are significantly smaller than those of the simple BP and SSA-BP models. This indicates that the Sine-SSA-BP model can achieve high-precision 4D trajectory prediction. The accuracy of trajectory prediction is notably improved by the sparrow search algorithm optimized with sine chaotic mapping, leading to faster convergence and better prediction outcomes, which better meet the requirements of aviation safety and control.

Enhanced Force Sensing Method Using BP Neural Network and Sensitive Point Analysis for Forging Manipulator Clamping Mechanism in Operational Conditions

Given the challenging operational environment of forging machines, characterized by harsh conditions, dynamic load fluctuations, and multi-axis force coupling, traditional force measurement and sensing methods often fail to meet the high demands for precision and reliable load sensing. To address this challenge, this study introduces a novel force sensing methodology based on a Backpropagation (BP) neural network for the clamping mechanism of forging machines. First, a force sensing model is developed to describe the behavior of the clamping mechanism under forging conditions, elucidating the force transmission laws and identifying key hinge points for force sensing. Next, a BP neural network model is constructed to analyze and decouple the forging state, exploring the impact of hidden layer neuron nodes on decoupling effectiveness and determining the optimal configuration of these nodes. The effectiveness of the proposed methodology is demonstrated through the network’s ability to decouple both the direction and magnitude of internal forging forces. The results confirm that this approach enables accurate and reliable force sensing for the clamping mechanism, providing a robust foundation for force control in forging machines.

Unveiling the power of R: a comprehensive perspective for laboratory medicine data analysis

Abstract R language has gained traction in laboratory medicine for its statistical power and dynamic tools like RMarkdown and RShiny. However, there is limited literature summarizing R packages and functions tailored for laboratory medicine, making it difficult for clinical laboratory workers to access these tools. Additionally, varying algorithms across R packages can lead to inconsistencies in published reports. This review addresses these challenges by providing an overview of R’s evolution and its key features, followed by a summary of statistical methods implemented in R, including platform comparisons, precision verification, factor analysis, and the establishment of reference intervals (RIs). We also highlight the development and validation of predictive models using techniques such as linear and logistic regression, decision trees, random forests, support vector machines, naive Bayes, K-Nearest Neighbors, k-means clustering, and backpropagation neural networks – all implemented in R. To ensure transparency and reproducibility in research, a checklist is provided for authors publishing papers using R for data analysis in laboratory medicine. In the final section, the potential of R in big data analytics is explored, focusing on standardized reporting through RMarkdown and the creation of user-friendly data visualization platforms with RShiny. Moreover, the integration of large language models (LLMs), such as ChatGPT, is discussed for their benefits in enhancing R programming, automating reporting, and offering insights from data analysis, thus improving the efficiency and accuracy of laboratory data analysis.

Logging-data-driven lithology identification in complex reservoirs: an example from the Niuxintuo block of the Liaohe oilfield

For lithologic oil reservoirs, lithology identification plays a significant guiding role in exploration targeting, reservoir evaluation, well network adjustment and optimization, and the establishment of reservoir models. Lithology is usually predicted from well log data based on limited core observations. In recent years, machine learning algorithms have been applied to lithology identification to enhance prediction accuracy. In this paper, five algorithms, including Bayes discriminant analysis, Random Forest (RF), Support Vector Machine (SVM), Back Propagation Neural Network (BPNN), and Convolutional Neural Network (CNN) are evaluated for lithology identification using data from the Niuxintuo reservoir. This reservoir is characterized by complex structural and sedimentary features, strong heterogeneity, and intricate lithological properties, all of which present considerable challenges for well logging identification. First, we conducted a detailed observation of the core lithology. Based on the requirements for reservoir evaluation and the principles of logging identification, we reclassify the lithology of the study area into two categories: clastic rocks and dolomite. The clastic rocks are further subdivided into five rock types: fine sandstone, medium-coarse sandstone, conglomerate, mudstone, and transitional rock. The well log series were selected through sensitivity analysis. Then, Bayes discriminant analysis and four machine learning methods were trained to identify the lithology of the study area. The results indicate that except for Bayes discriminant analysis, all the constructed machine learning classifiers demonstrate high prediction accuracy, with the accuracy rate exceeding 85%. Among them, SVM classifier shows the best performance achieving a prediction accuracy as high as 93%. Additionally, the well-trained SVM model was successfully used to predict the lithology profile of blind wells. Our findings provide valuable guidance for predicting the remaining oil distribution and further exploration potential in the Niuxintuo oilfield. Furthermore, this study gains insight into the process and methodology of rapidly predicting lithology of hydrocarbon reservoirs using easily accessible well logging data.

Prediction of temperature and humidity in footwear warehouses based on improved arithmetic optimization twin support vector machine for regression

Abstract Temperature and humidity fluctuations in storage environments significantly impact the material properties and quality of footwear. Temperature and humidity fluctuations in storage environments significantly impact the material properties and quality of footwear. Traditional warehouse management approaches typically rely on fixed environmental control systems, which struggle to adapt to real-time dynamic changes, leading to high energy consumption, inefficient management, and potential impacts on footwear quality. To address the aforementioned issues, the study employed an improved arithmetic optimization algorithm (AOA) to optimize the twin support vector machine for regression algorithm (LAOA-TSVR) and develop the corresponding prediction model. First, SPSS was used for correlation analysis to identify the primary factors influencing temperature and humidity in the warehouse. Second, to validate the model’s prediction accuracy, the algorithm was compared with three common models: backpropagation neural network (BP), radial basis function neural network (RBF), and twin support vector regression (TSVR). The results demonstrate that the LAOA-TSVR exhibits fast convergence, high prediction accuracy, and robust generalization ability. Finally, the model was applied in a real warehouse environment. The results indicate that when the temperature and humidity errors in the warehouse remain within ±10% and ±5%, the endpoint hit rates for warehouse temperature reach 96.8% and 88.1%. while those for warehouse humidity reach 93.6% and 86.3%, respectively. For ±10% errors, the dual hit rate for warehouse temperature and humidity is 92.7%. The model establishes a theoretical foundation for accurately predicting temperature and humidity in warehouse environments, thereby facilitating the optimization of control strategies and improving warehouse management efficiency.

Adaptive hybrid optimization for backpropagation neural networks in image classification

Image classification is essential in artificial intelligence, with applications in medical diagnostics, autonomous navigation, and industrial automation. Traditional training methods like stochastic gradient descent (SGD) often suffer from slow convergence and local minima. This research presents a hybrid Particle Swarm Optimization (PSO)-Genetic Algorithm (GA)-Backpropagation framework to enhance neural network training. By integrating AdaGrad and PSO for weight optimization, GA for refinement, and backpropagation for fine-tuning, the model improves performance. Results show a 97.5% accuracy on MNIST, a 5% improvement over Adam, and 40% faster convergence than SGD. This approach enhances efficiency, accuracy, and generalization, making it valuable for high-dimensional AI tasks.

A 3D optical microscopic imaging system based on PSO-BPNN-PID control and image pyramid transform fusion with experimental verification

Abstract The limited depth of field of optical microscopes typically results in the blurring of all but a small portion of the image, making it difficult to assess the height, shape, and surface profile of samples. This issue can be partly resolved by capturing a series of images at various sample distances by adjusting the piezoelectric stage; however, scanning performance is prone to perturbations resulting from nonlinear piezoelectric response characteristics. This paper presents a novel three-dimensional optical microscopic imaging system that uses a particle swarm optimization control scheme in parallel with a back-propagation neural network to achieve high-precision trajectory tracking. The system also features an algorithm to perform image pyramid transform fusion, image stitching, and 2D-to-3D conversion to enable the capture of 3D surface profiles of high accuracy. In experiments, the proposed scheme significantly outperformed existing methods in terms of accuracy and the level of detail.

Study on the dynamic prediction and optimum regulation scheme of water resource carrying capacity in the yellow river basin

The dynamic change in regional water resource carrying capacity (WRCC) is an important factor in formulating an optimal water resource allocation scheme. To accurately predict the dynamic change in WRCC and formulate the optimal regulation scheme in a river basin, this paper proposes a diagnostic index system of WRCC based on the mutual feeding mechanism of economic-society, water resources and eco-environment. The multi-methods including support vector machine (SV) and system dynamics (SD), back propagation neural network (BP), etc., were used to construct a quantitative dynamic prediction model for the WRCC. The orthogonal test (OT) method is used to optimize the WRCC regulation scheme. Through simulation, the WRCC in the sub-regions of the Yellow River Basin in 2030, 2035 and 2050 under different scenarios are obtained. The study results show that the WRCC in the upper, middle and lower reaches of the Yellow River Basin are overloaded or seriously overloaded before the operation of the Western Line of South-to-North Water Transfer Project (WLSNWTP). Six key driving indexes (unconventional water utilization, water flow into the sea, irrigation water use, irrigation area, industrial added value and water consumption of 10 thousand Yuan (CNY) industrial added value) are selected to carry out orthogonal tests, and the optimal WRCC regulation scheme is obtained in different level years. The present study results have an important reference value for efficient water resource utilization and eco-society development in the Yellow River Basin.

Integrating Functional Principal Component Analysis with Generalized Regression Neural Networks for Enhanced Forecasting of New Energy Vehicle Market Trends

With the rapid development of the new energy vehicle (NEV) industry, it has become a critical component of modern productivity, offering immense potential for economic growth and environmental benefits. However, accurately predicting NEV trends, such as sales volume and the annual installed capacity of power batteries, remains challenging due to the high dimensionality of data and the limitations of existing predictive methods. In view of these challenges, this paper proposes a novel prediction method that combines Functional Principal Component Analysis (FPCA) and Generalized Regression Neural Network (GRNN) to forecast NEV sales and power battery installation trends. By employing FPCA to reduce the dimensionality of nine key variables and using GRNN to incorporate seven influencing factors, such as ownership and total energy consumption, the study constructs a robust predictive model. The GRNN model is further optimized using cross-validation, achieving a high level of accuracy. It is hoped that the proposed FPCA-GRNN method, which has demonstrated superior performance compared to traditional approaches such as BP neural networks and multiple linear regression, will serve as a valuable tool for predicting NEV development trends and provide guidance for industry growth and policy-making in the NEV sector.

Research on the application of particle swarm optimization algorithm based on heterogeneous lithology analysis and drilling parameter correction to optimize the back propagation neural network rate of penetration prediction model

Accurate prediction of the rate of penetration (ROP) is of great significance for efficient drilling. The traditional ROP prediction model mostly ignores the influence of formation, and takes the term estimated downhole weight on bit (TTEDWOB) as the real downhole weight on bit (WOB) input model, which has a large prediction error. In this study, we modified TTEDWOB, and then proposed a back propagation (BP) neural network ROP prediction model based on particle swarm optimization (PSO) on the basis of formation classification. The results show that the error of ROP prediction model is reduced by 7.7% on average in different formation experiments after the correction of TTEDWOB; The prediction accuracy of the model after formation classification is increased by 9.5%; The model is applied to two wells in different blocks, and the average relative errors are 8.51% and 24.78% respectively. The difference between our work and the previous work is that we consider the correction of TTEDWOB and formation conditions. Our suggestion is that it is necessary for researchers to consider TTEDWOB correction and formation factors when establishing ROP prediction models, because doing so will help improve the prediction accuracy of the model.

Monitoring Canopy Height in the Hainan Tropical Rainforest Using Machine Learning and Multi-Modal Data Fusion

Biomass carbon sequestration and sink capacities of tropical rainforests are vital for addressing climate change. However, canopy height must be accurately estimated to determine carbon sink potential and implement effective forest management. Four advanced machine-learning algorithms—random forest (RF), gradient boosting decision tree, convolutional neural network, and backpropagation neural network—were compared in terms of forest canopy height in the Hainan Tropical Rainforest National Park. A total of 140 field survey plots and 315 unmanned aerial vehicle photogrammetry plots, along with multi-modal remote sensing datasets (including GEDI and ICESat-2 satellite-carried LiDAR data, Landsat images, and environmental information) were used to validate forest canopy height from 2003 to 2023. The results showed that RH80 was the optimal choice for the prediction model regarding percentile selection, and the RF algorithm exhibited the optimal performance in terms of accuracy and stability, with R2 values of 0.71 and 0.60 for the training and testing sets, respectively, and a relative root mean square error of 21.36%. The RH80 percentile model using the RF algorithm was employed to estimate the forest canopy height distribution in the Hainan Tropical Rainforest National Park from 2003 to 2023, and the canopy heights of five forest types (tropical lowland rainforests, tropical montane cloud forests, tropical seasonal rainforests, tropical montane rainforests, and tropical coniferous forests) were calculated. The study found that from 2003 to 2023, the canopy height in the Hainan Tropical Rainforest National Park showed an overall increasing trend, ranging from 2.95 to 22.02 m. The tropical montane cloud forest had the highest average canopy height, while the tropical seasonal forest exhibited the fastest growth. The findings provide valuable insights for a deeper understanding of the growth dynamics of tropical rainforests.

Frost resistance prediction of recycled aggregate concrete based on supervised algorithms

The frost resistance of recycled aggregate concrete (RAC) is a crucial performance indicator in its engineering application. Considering the complexity of factors that affect the frost resistance of RAC, in this study five neural network algorithms were employed, back-propagation neural network, radial basis function network, convolutional neural network, support vector machine and random forest regressor, along with two optimisations methods, particle swarm optimisation-back-propagation (PSO-BPNN) and genetic algorithm-back-propagation to predict the frost resistance of RAC. A database was compiled from 616 mixes in the peer-reviewed literature, and eight variables affecting the freeze–thaw resistance of RAC were taken as inputs. The compressive strength, mass loss rate and relative dynamic elastic modulus after freeze–thaw cycles were taken as outputs. The results showed that the main factors affecting the freezing resistance of RAC were the number of freeze–thaw cycles, the replacement rate of recycled aggregates, crushing index and water–cement ratio. Among the seven algorithms, the PSO-BPNN model had the best comprehensive prediction performance, with R2 of predicted compressive strength reaching 0.9714. It provides a reference value for further research on RAC frost resistance.