Neural Network-based Prediction Research Articles

Conformational Variability Prediction of Influenza Virus Hemagglutinins with Amino Acid Mutations Using Supersecondary Structure Code.

Supersecondary structure code (SSSC), which is represented as a conformation term using the letters "H," "S," "T," and "D" for each amino acid peptide unit, can be utilized to look up supersecondary structure motifs like a helix-hairpin-helix (HhH) motif as character strings from the Protein Data Bank (PDB) structure files. The deep neural network-based conformational variability prediction system of protein structures (SSSCPreds) can simultaneously predict locations of protein flexibility or rigidity and the shapes of those regions with high accuracy. The sequence flexibility/rigidity map obtained from SSSCPreds has the prediction accuracy enough to discuss the correlation with the sequence-to-phenotype ones by mutations. In this chapter, the protocol of conformational variability prediction methods using SSSC, including the analysis of influenza virus hemagglutinins with amino acid mutations, is described. The conformational variability pattern of hemagglutinins for human influenza A H1N1pdm extremely resembles that of avian influenza A H5N1, except for the furin cleavage site of H5N1. The transition of virus variants is visually understandable from the maps, including the sharply increased flexibility with the insight into the recent unseasonal influenza epidemics in Japan. The prediction accuracy of conformational variability is low for proteins at pH5 with very few measurement conditions, so this is the limitation of the conformational variability prediction method.

Feed-Forward Neural Network-Based Prediction of Flutter Speed of 3 DOF 2D Wing Model

IntroductionThe application of artificial intelligence, particularly neural network models, has expanded throughout various fields of academic and industrial studies, recently. This research focuses to investigate aeroelastic phenomena and predict flutter speeds without expensive computational and experimental methods utilizing these algorithms, specifically the artificial neural network (ANN).MethodsFundamentally, the approach involves an ANN algorithm to navigate the complexities of aeroelastic systems, achieved by layering multiple ANNs. The network's neurons effectively interpret diverse flight data by this structure. The study also incorporates state-space representation and Theodorsen's unsteady aerodynamic theory to create a comprehensive dataset on aeroelastic flutter speed across different wing parameters. In order to forecast the flutter speeds, a feed-forward neural network (FFNN) model, which is an approach of ANN method, containing sigmoid hidden neurons and a linear output neuron has been proposed for the conditions above.ResultsThe proposed model accomplished a regression value of 0.92 for flutter speeds according to the experimental findings. As presented, the suggested FFNN model is successful and can be employed to predict the flutter speed estimation. The findings of the FFNN structure are validated with the previous work for aeroelastic flutter speed analysis in the literature.DiscussionFinally, the findings indicate that the FFNN model formulated in this study is highly accurate in predicting flutter speeds with the accordance of numerical results of aeroelastic structure modeled with unsteady aerodynamic theory. In addition, the correctness of the predicted flutter speeds demonstrates that analyzing without expensive tests and high computational costs is in the near future.

Customer churn prediction model based on hybrid neural networks

In today’s competitive market environment, accurately identifying potential churn customers and taking effective retention measures are crucial for improving customer retention and ensuring the sustainable development of an organization. However, traditional machine learning algorithms and single deep learning models have limitations in extracting complex nonlinear and time-series features, resulting in unsatisfactory prediction results. To address this problem, this study proposes a hybrid neural network-based customer churn prediction model, CCP-Net. In the data preprocessing stage, the ADASYN sampling algorithm balances the sample sizes of churned and non-churned customers to eliminate the negative impact of sample imbalance on the model performance. In the feature extraction stage, CCP-Net uses Multi-Head Self-Attention to learn the global dependencies of the input sequences, combines with BiLSTM to capture the long-term dependencies in the sequential data, and uses CNN to extract the local features, and ultimately generates the prediction results. Experimental results of cross-validation on Telecom, Bank, Insurance, and News datasets show that CCP-Net outperforms the comparison algorithms in all performance metrics. For example, CCP-Net achieves a Precision of 92.19% on the Telecom dataset, 91.96% on the Bank dataset, 95.87% on the Insurance dataset, and 95.12% on the News dataset, which compares to other hybrid neural network models, the performance improvement of CCP-Net ranges from 1% to 3%. These results indicate that the design of the CCP-Net model effectively improves the accuracy and robustness of churn prediction, enabling it to be widely applied to different industries, especially in the financial, telecommunication, and media fields, to provide more comprehensive and effective churn management strategies for enterprises.

Artificial Neural Network-Based Prediction and Optimization of Polymer Membrane for Alkaline Water Electrolysis

Artificial Neural Network-Based Prediction and Optimization of Polymer Membrane for Alkaline Water Electrolysis

Hybrid Preprocessing for Neural Network-Based Stock Price Prediction

In the domain of stock price prediction, the intricate interdependencies within multivariate time series data present significant challenges for accurate forecasting. This paper introduces a groundbreaking hybrid preprocessing technique to tackle this issue. By leveraging the Empirical Wavelet Transform (EWT), we adeptly extract both low-frequency and high-frequency components from the time series. We then apply Dynamic Time Warping (DTW) and Differential Dynamic Time Warping (DDTW) to measure component similarities, identifying correlated patterns within the stock price series. High-frequency components are managed using sliding windows and Principal Component Analysis (PCA), while PCA is directly applied to low-frequency components. Integrating these techniques into neural network models, our approach yields a substantial 30% improvement in prediction accuracy compared to traditional methods. This significant advancement underscores the potential of our hybrid preprocessing method in enhancing stock price prediction accuracy, offering valuable insights for financial market analysis.

Neural Network-Based Prediction of Wind-Induced Responses in Large-Span Air-Supported Membrane Structures

Neural Network-Based Prediction of Wind-Induced Responses in Large-Span Air-Supported Membrane Structures

General neural network-based static performance prediction model construction techniques for gate-all-around and planar field effect transistor

—This paper proposes general neural network-based static performance prediction model construction techniques for gate-all-around (GAA) and planar field effect transistor (FET). Firstly, a unique data preprocessing method named quasi-linear transformation is proposed to improve the prediction accuracy. By introducing transformation functions to process the input, the relationship between the input and output is simplified, thereby facilitating the model training. Secondly, an improved weighted loss function scheme that considers a more comprehensive evaluation criterion to enhance the training process is proposed. Compared with traditional artificial neural networks, the average prediction error of the output and transfer curves is reduced by 33 % and 25 % for GAA and planar FET, respectively. Meanwhile, the proposed model demonstrates strong extrapolation ability. Moreover, compared to traditional methods of obtaining static characteristic curves, this method is more efficient. Furthermore. the proposed neural network-based static performance prediction model is converted to Verilog-A model, demonstrating potential in circuit simulation.

Vehicle Interactive Dynamic Graph Neural Network-Based Trajectory Prediction for Internet of Vehicles

Vehicle Interactive Dynamic Graph Neural Network-Based Trajectory Prediction for Internet of Vehicles

Neural network-based predictions of antimicrobial resistance phenotypes in multidrug-resistant Acinetobacter baumannii from whole genome sequencing and gene expression.

Whole genome sequencing (WGS) potentially represents a rapid approach for antimicrobial resistance genotype-to-phenotype prediction. However, the challenge still exists to predict fully minimum inhibitory concentrations (MICs) and antimicrobial susceptibility phenotypes based on WGS data. This study aimed to establish an artificial intelligence-based computational approach in predicting antimicrobial susceptibilities of multidrug-resistant Acinetobacter baumannii from WGS and gene expression data. Antimicrobial susceptibility testing (AST) was performed using the broth microdilution method for 10 antimicrobial agents. In silico multilocus sequence typing (MLST), antimicrobial resistance genes, and phylogeny based on cgSNP and cgMLST strategies were analyzed. High-throughput qPCR was performed to measure the expression level of antimicrobial resistance (AMR) genes. Most isolates exhibited a high level of resistance to most of the tested antimicrobial agents, with the majority belonging to the IC2/CC92 lineage. Phylogenetic analysis revealed undetected transmission events or local outbreaks. The percentage agreements between AMR phenotype and genotype ranged from 70.08% to 89.96%, with the coefficient of agreement (κ) extending from 0.025 and 0.881. The prediction of AST employed by deep neural network models achieved an accuracy of up to 98.64% on the testing data set. Additionally, several linear regression models demonstrated high prediction accuracy, reaching up to 86.15% within an error range of one gradient, indicating a linear relationship between certain gene expressions and the corresponding antimicrobial MICs. In conclusion, neural network-based predictions could be used as a tool for the surveillance of antimicrobial resistance in multidrug-resistant A. baumannii.

Postoperative Karnofsky performance status prediction in patients with IDH wild-type glioblastoma: A multimodal approach integrating clinical and deep imaging features.

Glioblastoma is a highly aggressive brain tumor with limited survival that poses challenges in predicting patient outcomes. The Karnofsky Performance Status (KPS) score is a valuable tool for assessing patient functionality and contributes to the stratification of patients with poor prognoses. This study aimed to develop a 6-month postoperative KPS prediction model by combining clinical data with deep learning-based image features from pre- and postoperative MRI scans, offering enhanced personalized care for glioblastoma patients. Using 1,476 MRI datasets from the Brain Tumor Segmentation Challenge 2020 public database, we pretrained two variational autoencoders (VAEs). Imaging features from the latent spaces of the VAEs were used for KPS prediction. Neural network-based KPS prediction models were developed to predict scores below 70 at 6 months postoperatively. In this retrospective single-center analysis, we incorporated clinical parameters and pre- and postoperative MRI images from 150 newly diagnosed IDH wild-type glioblastoma, divided into training (100 patients) and test (50 patients) sets. In training set, the performance of these models was evaluated using the area under the curve (AUC), calculated through fivefold cross-validation repeated 10 times. The final evaluation of the developed models assessed in the test set. Among the 150 patients, 61 had 6-month postoperative KPS scores below 70 and 89 scored 70 or higher. We developed three models: a clinical-based model, an MRI-based model, and a multimodal model that incorporated both clinical parameters and MRI features. In the training set, the mean AUC was 0.785±0.051 for the multimodal model, which was significantly higher than the AUCs of the clinical-based model (0.716±0.059, P = 0.038) using only clinical parameters and the MRI-based model (0.651±0.028, P<0.001) using only MRI features. In the test set, the multimodal model achieved an AUC of 0.810, outperforming the clinical-based (0.670) and MRI-based (0.650) models. The integration of MRI features extracted from VAEs with clinical parameters in the multimodal model substantially enhanced KPS prediction performance. This approach has the potential to improve prognostic prediction, paving the way for more personalized and effective treatments for patients with glioblastoma.

Edge Computing-driven Smart Transportation: Analysis and Prediction of Urban Traffic Flow

Abstract With the continuous development of artificial intelligence, the research and application of automated connected driving is receiving more and more attention. The combination of mobile edge computing and Telematics provides crucial technical support. In this paper, we propose a rasterization method for mobile edge computing scenario. We establish a grid classification model based on K-means to facilitate targeted and distributed optimization of computing resources. Meanwhile, we dynamically adjust the allocation of computational resources and the location of edge servers in advance deployment by predicting vehicle traffic. The experimental results demonstrate that the K-means-based grid classification model can accurately classify the grids and differentiate the traffic patterns of the grids, such as remote suburban areas, suburban areas, residential living areas, production industrial areas, and commercial center areas. The Backpropagation (BP) neural network-based grid traffic flow prediction model can simulate the overall changes of the grid traffic flow throughout a day to the maximum extent. Finally, we plan optimization scheme in urban mobile edge scenarios based on the results obtained from our analysis and prediction, and through our results, we can make edge computing resources be utilized efficiently.

Optimizing prediction accuracy in dynamic systems through neural network integration with Kalman and alpha-beta filters.

In the realm of dynamic system analysis, the Kalman filter and the alpha-beta filter are widely recognized for their tracking and prediction capabilities. However, their performance is often limited by static parameters that cannot adapt to changing conditions. Addressing this limitation, this paper introduces innovative neural network-based prediction models that enhance the adaptability and accuracy of these conventional filters. Our approach involves the integration of neural networks within the filtering algorithms, enabling the dynamic augmentation of parameters in response to performance feedback. We present two modified filters: a neural network-based Kalman filter and an alpha-beta filter, each augmented to incorporate neural network-driven parameter tuning. The alpha-beta filter is enhanced with neural network outputs for its α and β parameters, while the Kalman filter employs a neural network to optimize its internal parameter R and noise factor F. We assess these advanced models using the root mean square error (RMSE) metric, where our neural network-based alpha-beta filter demonstrates a significant 38.2% improvement in prediction accuracy, and the neural network-based Kalman filter achieves a 53.4% enhancement. Hence, our novel approach of integrating neural networks into filtering algorithms stands out as an effective strategy to significantly enhance their performance in dynamic environments.

Artificial neural network-based prediction of multiple sclerosis using blood-based metabolomics data

Multiple sclerosis (MS) remains a challenging neurological condition for diagnosis and management and is often detected in late stages, delaying treatment. Artificial intelligence (AI) is emerging as a promising approach to extracting MS information when applied to different patient datasets. Given the critical role of metabolites in MS profiling, metabolomics data may be an ideal platform for the application of AI to predict disease. In the present study, a machine-learning (ML) approach was used for a detailed analysis of metabolite profiles and related pathways in patients with MS and healthy controls (HC). This approach identified unique alterations in biochemical metabolites and their correlation with disease severity parameters. To enhance the efficiency of using metabolic profiles to determine disease severity or the presence of MS, we trained an AI model on a large volume of blood-based metabolomics datasets. We constructed this model using an artificial neural network (ANN) architecture with perceptrons. Data were divided into training, validation, and testing sets to determine model accuracy. After training, accuracy reached 87 %, sensitivity was 82.5 %, specificity was 89 %, and precision was 77.3 %. Thus, the developed model seems highly robust, generalizable with a wide scope and can handle large amounts of data, which could potentially assist neurologists. However, a large multicenter cohort study is necessary for further validation of large-scale datasets to allow the integration of AI in clinical settings for accurate diagnosis and improved MS management.

A neural network model for predicting the effectiveness of treatment in patients with neovascular glaucoma associated with diabetes mellitus.

The study hypothesizes that neural networks can be an effective tool for predicting treatment outcomes in patients with diabetic neovascular glaucoma (NVG), considering not only baseline intraocular pressure (IOP) values but also inflammation and intraocular microcirculation indicators. To investigate the diagnostic significance of inflammation and intraocular blood circulation indicators in a neural network model predicting the effectiveness of transscleral cyclophotocoagulation (TSC CPC) treatment in patients with NVG of diabetic origin. This retrospective cohort study included 127 patients (127 eyes; aged Me 65.0 years) with painful diabetic NVG and 20 healthy individuals (aged Me 61.5 years) as an immunological control. All patients underwent TSC CPC with a diode laser. Treatment success was defined as achieving an IOP level of ≤ 21 mmHg and maintaining or improving best-corrected visual acuity (BCVA) after 12 months of observation. Preoperative systemic immune-inflammation index (SII = platelets × [neutrophils/lymphocytes]) and systemic inflammation response index (SIRI = neutrophils × [monocytes/lymphocytes]) were calculated. We assessed the values of volumetric pulse blood filling, determined by the rheographic coefficient (RQ, 0/00), using the rheoophthalmography (ROG) method. Multiple regression analysis was used to conclude the significance of treatment efficacy based on initial clinical and laboratory indicators, followed by constructing a prediction model in the neural network. The development of the neural network model identified the most significant "input" parameters: SIRI (100%), RQ (85.7%), and SII (80.7%), which significantly influenced treatment success. The sensitivity of the neural network model was 100%, specificity was 30%, and the percentage of correctly predicted events during testing on the control group was 92.9%. Neural network-based prediction of transscleral cyclophotocoagulation effectiveness for patients with diabetic neovascular glaucoma allows for a sufficiently accurate forecast of treatment success with a probability of 92.9%. We believe the in-time correction of systemic inflammation and intraocular blood circulation can significantly reduce intraocular pressure, preserve visual acuity, and improve the quality of life in patients with diabetic NVG after TSC CPC. Further research is required to support these findings.

Graph Neural Network-Based Molecular Property Prediction with Patch Aggregation.

Graph neural networks (GNNs) have emerged as powerful tools for quantum chemical property prediction, leveraging the inherent graph structure of molecular systems. GNNs depend on an edge-to-node aggregation mechanism for combining edge representations into node representations. Unfortunately, existing learnable edge-to-node aggregation methods substantially increase the number of parameters and, thus, the computational cost relative to simple sum aggregation. Worse, as we report here, they often fail to improve predictive accuracy. We therefore propose a novel learnable edge-to-node aggregation mechanism that aims to improve the accuracy and parameter efficiency of GNNs in predicting molecular properties. The new mechanism, called "patch aggregation", is inspired by the Multi-Head Attention and Mixture of Experts machine learning techniques. We have incorporated the patch aggregation method into the specialized, state-of-the-art GNN models SchNet, DimeNet++, SphereNet, TensorNet, and VisNet and show that patch aggregation consistently outperforms existing learnable and nonlearnable aggregation techniques (sum, multilayer perceptron, softmax, and set transformer aggregation) in the prediction of molecular properties such as QM9 thermodynamic properties and MD17 molecular dynamics trajectory energies and forces. We also find that patch aggregation not only improves prediction accuracy but also is parameter-efficient, making it an attractive option for practical applications for which computational resources are limited. Further, we show that Patch aggregation can be applied across different GNN models. Overall, Patch aggregation is a powerful edge-to-node aggregation mechanism that improves the accuracy of molecular property predictions by GNNs.

Towards artificial intelligence-based disease prediction algorithms that comprehensively leverage and continuously learn from real-world clinical tabular data systems.

This manuscript presents a proof-of-concept for a generalizable strategy, the full algorithm, designed to estimate disease risk using real-world clinical tabular data systems, such as electronic health records (EHR) or claims databases. By integrating classic statistical methods and modern artificial intelligence techniques, this strategy automates the production of a disease prediction model that comprehensively reflects the dynamics contained within the underlying data system. Specifically, the full algorithm parses through every facet of the data (e.g., encounters, diagnoses, procedures, medications, labs, chief complaints, flowsheets, vital signs, demographics, etc.), selects which factors to retain as predictor variables by evaluating the data empirically against statistical criteria, structures and formats the retained data into time-series, trains a neural network-based prediction model, then subsequently applies this model to current patients to generate risk estimates. A distinguishing feature of the proposed strategy is that it produces a self-adaptive prediction system, capable of evolving the prediction mechanism in response to changes within the data: as newly collected data expand/modify the dataset organically, the prediction mechanism automatically evolves to reflect these changes. Moreover, the full algorithm operates without the need for a-priori data curation and aims to harness all informative risk and protective factors within the real-world data. This stands in contrast to traditional approaches, which often rely on highly curated datasets and domain expertise to build static prediction models based solely on well-known risk factors. As a proof-of-concept, we codified the full algorithm and tasked it with estimating 12-month risk of initial stroke or myocardial infarction using our hospital's real-world EHR. A 66-month pseudo-prospective validation was conducted using records from 558,105 patients spanning April 2015 to September 2023, totalling 3,424,060 patient-months. Area under the receiver operating characteristic curve (AUROC) values ranged from .830 to .909, with an improving trend over time. Odds ratios describing model precision for patients 1-100 and 101-200 (when ranked by estimated risk) ranged from 15.3 to 48.1 and 7.2 to 45.0, respectively, with both groups showing improving trends over time. Findings suggest the feasibility of developing high-performing disease risk calculators in the proposed manner.

Performance degradation evaluation of square steel tubes with local penetration damage

Steel tubular structures are susceptible to welding defects, cracking defects and environmental corrosion, which can negatively affect the stability and reliability of structure. This study tested the axial compression performance of twenty-six square steel tubular stubs with localized penetration damage simulated by artificial notch, considering the influence of notch length, orientation, and position. The test results showed that the failure mode of the specimens was dominated by the local buckling along the direction of the notch axis. The failure mode of the specimen with vertical notch showed notch opening, while those with horizontal notch showed notch closing, resulting in the secondary peak value in the load-displacement curve. The existence of local penetration damage reduced the ultimate strength and ductility of the tubular stubs. Compared to the ultimate strength of the specimens with middle notch, those with corner notch were weakened more, up to 27 %. Furthermore, the effects of notch dimension, position, orientation, and tube wall thickness on the axial bearing capacity of square steel tubular stubs are discussed by using a validated finite element model. Owing to the analysis, the ultimate strength of the stubs is reduced by increasing notch length, width and depth, whereas the reduction effect tends to be weakened as the notch orientation angle increases. The notch located in the corner is the most unfavorable for the axial compression performance of the square steel tubular stubs. A semi-empirical formula for evaluating the ultimate strength of square steel tubular stubs with local penetration damage is proposed by introducing the strength reduction factor. An artificial neural network-based ultimate strength prediction model is developed that can consider the effects of wide-range parameters, achieving high accuracy.

A predictive model for classifying college students' academic performance based on visual-spatial skills.

As the application of visual-spatial skills in academic disciplines, vocational fields and daily life is becoming more and more prominent, it is of great theoretical and practical significance how to make use of big data and artificial intelligence technology to conduct research on the relationship between visual-spatial skills and students' grades. This paper explores and analyses from the perspective of artificial intelligence, combining students' visual-spatial skills and students' specific attribute characteristics to construct an expert system, which defines the prediction of academic performance as a classification problem corresponding to the five categories of excellent, good, moderate, passing, and weak, respectively, and based on which a deep neural network-based classification prediction model for students' performance is designed. The experimental results show that visual-spatial skills plays an important role in the professional learning of science and engineering students, while the classification model designed in this paper has high accuracy in the grade prediction process. This paper not only helps to fill the gaps in the current research field, but is also expected to provide scientific basis for educational practice and promote the development of the education field in a more intelligent and personalized direction.

Recurrent neural network-based prediction of O-GlcNAcylation sites in mammalian proteins

O-GlcNAcylation has the potential to be an important target for therapeutics, but a motif or an algorithm to reliably predict O-GlcNAcylation sites is not available. Current predictive models are insufficient as they fail to generalize, and many are no longer available. This article constructs recurrent neural network models to predict O-GlcNAcylation sites based on protein sequences. Different datasets are evaluated separately and assessed in terms of strengths and issues. Within a given dataset, results are robust to changes in cross-validation and test data as determined by nested validation. The best model achieves an F1 score of 36% (more than 3.5-fold greater than the previous best model) and a Matthews Correlation Coefficient of 35% (more than 4.5-fold greater than the previous best model), and, for the F1 score, 7.6-fold higher than when not using any model. Shapley values are used to interpret the model’s predictions and provide biological insight into O-GlcNAcylation.

Design Optimization of an Enhanced-Mode GaN HEMT with Hybrid Back Barrier and Breakdown Voltage Prediction Based on Neural Networks

To improve the breakdown voltage (BV), a GaN-based high-electron-mobility transistor with a hybrid AlGaN back barrier (HBB-HEMT) was proposed. The hybrid AlGaN back barrier was constructed using the Al0.25Ga0.75N region and Al0.1G0.9N region, each with a distinct Al composition. Simulation results of the HBB-HEMT demonstrated a breakdown voltage (1640 V) that was 212% higher than that of the conventional HEMT (Conv-HEMT) and a low on-resistance (0.4 mΩ·cm2). Ultimately, the device achieved a high Baliga’s figure of merit (7.3 GW/cm2) among reported devices of similar size. A back-propagation (BP) neural network-based prediction model was trained to predict BV for enhanced efficiency in subsequent work. The model was trained and calibrated, achieving a correlation coefficient (R) of 0.99 and a prediction accuracy of 95% on the test set. The results indicated that the BP neural network model using the Levenberg–Marquardt algorithm accurately predicted the forward breakdown voltage of the HBB-HEMT, underscoring the feasibility and significance of neural network models in designing GaN power devices.