Network-based Prediction Research Articles

Effects of Illuminance Level of Light Source on White Appearance of a Tablet Display

The appearance of white significantly impacts display image quality, requiring a neutral white point for optimal performance. This study explores how perceived whiteness changes under ambient illumination levels (150, 300, 600, and 1200 lx) and correlated color temperatures (3500 K and 6500 K). As a result, the adapted white points of light sources with different correlated color temperatures are similar at lower ambient illuminance levels. In comparison, their adaptation trends exhibit significant differences at higher illuminance levels. At 150 lx, adapted white points for 3500 K and 6500 K light sources shift toward higher color temperatures and converge. With increased illumination, the 3500 K white point shifts toward its light source, while the 6500 K white point shifts to a higher correlated color temperature. The neural network-based prediction model developed in this study accurately forecasts perceived whiteness across conditions, offering valuable design guidance for the display and lighting industries.

Deep Double Towers Click Through Rate Prediction Model with Multi-Head Bilinear Fusion

The click-through rate (CTR) forecast is among the mainstream research directions in the domain of recommender systems, especially in online advertising suggestions. Among them, the multilayer perceptron (MLP) has been extensively utilized as the cornerstone of deep CTR prediction models. However, current neural network-based CTR prediction models commonly employ a single MLP network to capture nonlinear interactions between high-order features, while disregarding the interaction among differentiated features, resulting in poor model performance. Although studies such as DeepFM have proposed dual-branch interaction models to learn complex features, they still fall short of achieving more nuanced feature fusion. To address these challenges, we propose a novel model, the Deep Double Towers model (DDT), which improves the accuracy of CTR prediction through multi-head bilinear fusion while incorporating symmetry in its architecture. Specifically, the DDT model leverages symmetric parallel MLP networks to capture the interactions between differentiated features in a more structured and balanced manner. Furthermore, the multi-head bilinear fusion layer enables refined feature fusion through symmetry-aware operations, ensuring that feature interactions are aligned and symmetrically integrated. Experimental results on publicly available datasets, such as Criteo and Avazu, show that DDT surpasses existing models in improving the accuracy of CTR prediction, with symmetry contributing to more effective and balanced feature fusion.

Preoperative comprehensive risk estimation for axillary lymph node metastasis in breast cancer: development and verification of a network-based prediction model

To prevent the overaggressive treatment of axillary lymph nodes (ALNs) in breast cancer, it is necessary to develop a convenient analysis method that accurately and comprehensively reflects whether ALNs are metastatic or nonmetastatic. We retrospectively analyzed data from patients who underwent surgery for breast cancer at the Weifang Hospital of Traditional Chinese Medicine between January 2019 and June 2023. Binary logistic regression analysis was used to predict the metastasis status of ALNs. The developmental data set included 531 patients (January 2019–June 2023). The validation set included separate data points (n = 178, January 2019–June 2023). Multivariate analysis revealed that positive findings on breast physical examination, ultrasound grades of ALNs, lymphovascular invasion, and Her-2 status had significant predictive value for metastatic ALNs. Based on these findings, a 5-grade risk scoring system and 3-level management recommendations were developed. The risk of metastasis ranged from 11.25 to 93.46%, which was positively correlated with an increase in risk grade. The areas under the curve of the development and validation sets were 0.895 and 0.865, respectively. Ultimately, a convenient, accurate and comprehensive web-based predictive model was constructed using various breast cancer clinical, imaging and pathological criteria to stratify ALNs according to the metastasis probability.

Enhancing Molecular Network-Based Cancer Driver Gene Prediction Using Machine Learning Approaches: Current Challenges and Opportunities.

Cancer is a complex disease driven by mutations in the genes that play critical roles in cellular processes. The identification of cancer driver genes is crucial for understanding tumorigenesis, developing targeted therapies and identifying rational drug targets. Experimental identification and validation of cancer driver genes are time-consuming and costly. Studies have demonstrated that interactions among genes are associated with similar phenotypes. Therefore, identifying cancer driver genes using molecular network-based approaches is necessary. Molecular network-based random walk-based approaches, which integrate mutation data with protein-protein interaction networks, have been widely employed in predicting cancer driver genes and demonstrated robust predictive potential. However, recent advancements in deep learning, particularly graph-based models, have provided novel opportunities for enhancing the prediction of cancer driver genes. This review aimed to comprehensively explore how machine learning methodologies, particularly network propagation, graph neural networks, autoencoders, graph embeddings, and attention mechanisms, improve the scalability and interpretability of molecular network-based cancer gene prediction.

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.

Contextual Deep Semantic Feature Driven Multi-Types Network Intrusion Detection System for IoT-Edge Networks

Recent years have witnessed an exponential rise in wireless networks and allied interoperable distributed computing frameworks, where the different sensory units transfer real-world event data to the network analyzer for run-time decisions. There exists an array of applications employing edge- internet of things (Edge-IoT) where the edge nodes collect local data to perform real-time decisions. However, the at-hand edge-IoT systems being decentralized, infrastructure-less, and dynamic remain vulnerable to man-in-the-middle attacks, intrusion, denial of service attacks, etc. Though in the past, numerous efforts were made towards intrusion detection in IoT networks, the major approaches focused merely on standalone intrusion detection, and therefore their scalability towards multiple attack detection remains unaddressed. On the contrary, applying a unit intrusion detection system for each type of attack can impose resource exhaustion and delay. Recently authors have used deep learning methods like convolutional neural network (CNN), and long- and short-term memory (LSTM) to perform learning-based intrusion detection. However, being reliant on merely local features its reliability remains suspicious. Such methods ignore long-term dependency problems that limit their efficacy in intrusion detection in temporal Edge-IoT network traffic. With this motivation, in this paper, a contextual deep semantic feature-driven multi-type intrusion detection model (CDS-MNIDS) is proposed for Edge-IoT networks. The proposed CDS-MNIDS model at first performs network traffic segmentation from the temporal network traces obtained from the network gateway. Subsequently, the node’s dynamic features including the node’s address, packet size, transmission behavior, etc., are processed for Word2Vec encoding, followed by a cascaded deep network-based learning and prediction. The CDS-MNIDS model embodied a cascaded deep network encompassing LSTM and bidirectional LSTM networks, where the first extracted local features. At the same time, the latter obtained contextual features from the input local feature vector. The extracted local and contextual features were projected to the global average pooling layer followed by the fully connected layer that in conjunction with the Softmax layer performed multi-class classification.

Neural Network-based Prediction and Optimization of Performance in Single Slope Solar Stills Enhanced with Nanoparticles for Improved Water Production

Neural Network-based Prediction and Optimization of Performance in Single Slope Solar Stills Enhanced with Nanoparticles for Improved Water Production

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

Single-microglia transcriptomic transition network-based prediction and real-world patient data validation identifies ketorolac as a repurposable drug for Alzheimer's disease.

High microglial heterogeneities hinder the development of microglia-targeted treatment for Alzheimer's disease (AD). We integrated 0.7 million single-nuclei RNA-sequencing transcriptomes from human brains using a variational autoencoder. We predicted AD-relevant microglial subtype-specific transition networks for disease-associated microglia (DAM), tau microglia, and neuroinflammation-like microglia (NIM). We prioritized drugs by specifically targeting microglia-specific transition networks and validated drugs using two independent real-world patient databases. We identified putative AD molecular drivers (e.g., SYK, CTSB, and INPP5D) in transition networks of DAM and NIM. Via specifically targeting NIM, we identified that usage of ketorolac was associated with reduced AD incidence in both MarketScan (hazard ratio [HR]=0.89) and INSIGHT (HR=0.83) Clinical Research Network databases, mechanistically supported by ketorolac-treated transcriptomic data from AD patient induced pluripotent stem cell-derived microglia. This study offers insights into the pathobiology of AD-relevant microglial subtypes and identifies ketorolac as a potential anti-inflammatory treatment for AD. An integrative analysis of≈ 0.7 million single-nuclei RNA-sequencing transcriptomes from human brains identified Alzheimer's disease (AD)-relevant microglia subtypes. Network-based analysis identified putative molecular drivers (e.g., SYK, CTSB, INPP5D) of transition networks between disease-associated microglia (DAM) and neuroinflammation-like microglia (NIM). Via network-based prediction and population-based validation, we identified that usage of ketorolac (a US Food and Drug Administration-approved anti-inflammatory medicine) was associated with reduced AD incidence in two independent patient databases. Mechanistic observation showed that ketorolac treatment downregulated the Type-I interferon signaling in patient induced pluripotent stem cell-derived microglia, mechanistically supporting its protective effects in real-world patient databases.

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.

Sparse attention regression network-based soil fertility prediction with UMMASO

The challenge of imbalanced soil nutrient datasets significantly hampers accurate predictions of soil fertility. To tackle this, a new method is suggested in this research, combining Uniform Manifold Approximation and Projection (UMAP) with Least Absolute Shrinkage and Selection Operator (LASSO). The main aim is to counter the impact of uneven data distribution and improve soil fertility models' predictive precision. The model introduced uses Sparse Attention Regression, effectively incorporating pertinent features from the imbalanced dataset. UMAP is utilised initially to reduce data complexity, unveiling hidden structures and essential patterns. Following this, LASSO is applied to refine features and enhance the model's interpretability. The experimental outcomes highlight the effectiveness of the UMAP and LASSO hybrid approach. The proposed model achieves outstanding performance metrics, reaching a predictive accuracy of 98 %, demonstrating its capability in accurate soil fertility predictions. It also showcases a Precision of 91.25 %, indicating its adeptness in accurately identifying fertile soil instances. The Recall metric stands at 90.90 %, emphasizing the model's ability to capture true positive cases effectively.

Multi-graph attention temporal convolutional network-based radius prediction in three-roller bending of thin-walled parts

Three-roller bending is the key production process for aerospace products, which forms thin-walled parts into a curved shape with a specific radius through multiple passes of these parts between loaded rollers. Radius prediction is of great importance for reasonable bending process control for desired curves. However, due to the various influence factors and the complicated interaction between successive passes, it prevents high accuracy of radius prediction. The prediction model based on multi-graph attention temporal convolutional network is therefore proposed to deal with these challenges. First, multiple graphs are constructed from multi-pass observations from three-roller bending process, with each graph representing the influencing factors of a specific pass. Second, graph attention mechanism explores the coupling effects of influence factors on the radius and realizes the extraction of key factors for each graph. Third, temporal convolutional network reveals the interaction between successive passes by establishing the connection between different graphs, and provides the radius prediction at each pass. In comparative experiments based on simulated data and experimental data collected from real cases, the results demonstrate the higher prediction accuracy of the proposed method over traditional methods.

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

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.

Characteristic Activation Pattern and Network Connectivity of Prefrontal Cortex in Patients with Type 2 Diabetes Mellitus and Major Depressive Disorder during a Verbal Fluency Task: A Functional Near-Infrared Spectroscopy Study Based on Network-Based Statistic Prediction.

Type 2 diabetes mellitus (T2DM) and major depressive disorder (MDD) together occur frequently among the elderly population. However, the inconsistency in assessments and limited medical resources in the community make it challenging to identify depression in patients with T2DM. This cross-sectional study aimed to investigate the activation pattern and network connectivity of prefrontal cortex (PFC) during a verbal fluency task (VFT) in patients with T2DM and MDD using functional near-infrared spectroscopy (fNIRS). Three parallel groups (T2DM with MDD group, T2DM group, and healthy group) with 100 participants in each group were included in the study. Recruitment took place from August 1, 2020, to December 31, 2023. Due to the close association between the PFC and depressive emotions, fNIRS was used to monitor brain activation and network connectivity of PFC in all participants during a task of Chinese-language phonological VFT. Network-based statistic prediction was adopted as data analysis method. Patients in the T2DM with MDD group showed characteristic activation pattern and network connectivity in contrast with patients with T2DM and healthy controls, including decreased activation in PFC, and decreased network connectivity of right dorsolateral prefrontal cortex (DLPFC). Furthermore, the network connectivity of the right DLPFC in patients with T2DM and MDD was negatively correlated with scores of Hamilton Depression Scale-24 (HAMD-24). There was a distinctive activation pattern and network connectivity of the PFC in patients with T2DM and MDD. The right DLPFC could serve as a potential target for the diagnosis and intervention of MDD in patients with T2DM.

Multi-scale self-attention feature decoupling transfer network-based cross-domain capacity prediction of lithium-ion batteries

Lithium-ion battery capacity prediction is paramount for improving the safety and reliability of energy storage devices. However, accurately predicting capacity continues to pose a challenge, stemming from the diverse range of operational conditions and the lack of capacity labels under unpredictable scenarios. To tackle this problem, a multi-scale self-attention feature decoupling transfer network is proposed for predicting the capacity of lithium-ion batteries across diverse operational conditions. More specifically, multi-scale multi-head self-attentive encoders are designed to adaptively extract multi-scale domain-invariant features by using a series of loss function constraints. Then, transfer learning integrated with domain-invariant features is applied to migrate the knowledge of the source domain to the target domain for cross-domain capacity prediction. Finally, the performance of our methods is authenticated through utilization of two battery dataset, each including three different charging and discharging scenarios. Experimental results indicate that the proposed method effectively extracts domain-invariant features from charging data characterized by varying distributions, thereby mitigating the effects of distributional differences. Compared with the state-of-the-art methods, the proposed method shows excellent accuracy and bustling ability in two datasets.