Construction Of Model Research Articles

Construction and Validation of Surface Soil Moisture Inversion Model Based on Remote Sensing and Neural Network

Surface soil moisture (SSM) reflects the dry and wet states of soil. Microwave remote sensing technology can accurately obtain regional SSM in real time and effectively improve the level of agricultural drought monitoring, and it is of great significance for agricultural precision irrigation and smart agriculture construction. Based on Sentinel-1, Sentinel-2, and Landsat-8 images, the effect of vegetation was removed by the water cloud model (WCM), and SSM was retrieved and validated by a radial basis function (RBF) neural network model in bare soil and vegetated areas, respectively. The normalized difference vegetation index (NDVI) calculated by Landsat-8 (NDVI_Landsat-8) had a better effect on removing the influence the of vegetation layer than that of NDVI_Sentinel-2. The RBF network model, established in a bare area (R = 0.796; RMSE = 0.029 cm3/cm3), and the RBF neural network model, established in vegetated areas (R = 0.855; RMSE = 0.024 cm3/cm3), have better simulation effects on SSM than a linear SSM inversion model with single polarization. The introduction of surface parameters to the RBF neural network model can improve the accuracy of the model and realize the high-accuracy inversion of SSM in the study area.

The statistical and dynamic modeling of the first part of the 2013-2014 Euromaidan protests in Ukraine: The Revolution of Dignity and preceding times.

Ukraine's tug-of-war between Russia and the West has had significant and lasting consequences for the country. In 2013, Viktor Yanukovych, the Ukrainian president aligned with Russia, opted against signing an association agreement with the European Union. This agreement aimed to facilitate trade and travel between the EU and Ukraine. This decision sparked widespread protests that coalesced in Kyiv's Maidan Square, eventually becoming known as the Euromaidan protests. In this study, we analyze the protest data from 2013, sourced from Ukraine's Center for Social and Labor Research. Despite the dataset's limitations and occasional inconsistencies, we demonstrate the extraction of valuable insights and the construction of a descriptive model from such data. Our investigation reveals a pre-existing state of self-excitation within the system even before the onset of the Euromaidan protests. This self-excitation intensified during the Euromaidan protests. A statistical analysis indicates that the government's utilization of force correlates with increased future protests, exacerbating rather than quelling the protest movement. Furthermore, we introduce the implementation of Hawkes process models to comprehend the spatiotemporal dynamics of the protest activity. Our findings highlight that, while protest activities spread across the entire country, the driving force behind the dynamics of these protests was the level of activity in Kyiv. Furthermore, in contrast to prior research that emphasized geographical proximity as a key predictor of event propagation, our study illustrates that the political alignment among oblasts, which are the distinct municipalities comprising Ukraine, had a more profound impact than mere geographic distance. This underscores the significance of social and cultural factors in molding the trajectory of political movements.

Mathematical Models of the Arabidopsis Circadian Oscillator

We review the construction and evolution of mathematical models of the Arabidopsis circadian clock, structuring the discussion into two distinct historical phases of modeling strategies: extension and reduction. The extension phase explores the bottom-up assembly of regulatory networks, introducing as many components and interactions as possible to capture the oscillatory nature of the clock. The reduction phase deals with functional decomposition, distilling complex models to their essential dynamical repertoire. Current challenges in this field, including the integration of spatial considerations and environmental influences like light and temperature, are also discussed. The review emphasizes the ongoing need for models that balance molecular detail with practical simplicity.

Detecting Road Intersections from Crowdsourced Trajectory Data Based on Improved YOLOv5 Model

In recent years, the rapid development of autonomous driving and intelligent driver assistance has brought about urgent demands on high-precision road maps. However, traditional road map production methods mainly rely on professional survey technologies, such as remote sensing and mobile mapping, which suffer from high costs, object occlusions, and long updating cycles. In the era of ubiquitous mapping, crowdsourced trajectory data offer a new and low-cost data resource for the production and updating of high-precision road maps. Meanwhile, as key nodes in the transportation network, maintaining the currency and integrity of road intersection data is the primary task in enhancing map updates. In this paper, we propose a novel approach for detecting road intersections based on crowdsourced trajectory data by introducing an attention mechanism and modifying the loss function in the YOLOv5 model. The proposed method encompasses two key steps of training data preparation and improved YOLOv5s model construction. Multi-scale training processing is first adopted to prepare a rich and diverse sample dataset, including various kinds and different sizes of road intersections. Particularly to enhance the model’s detection performance, we inserted convolutional attention mechanism modules into the original YOLOv5 and integrated other alternative confidence loss functions and localization loss functions. The experimental results demonstrate that the improved YOLOv5 model achieves detection accuracy, precision, and recall rates as high as 97.46%, 99.57%, and 97.87%, respectively, outperforming other object detection models.

Early warning model based on support vector machine ensemble algorithm

Support Vector Machine (SVM) is a powerful machine learning technique often applied in various early warning scenarios. When using SVM for binary classification, its output results can be converted into a probability distribution based on sample categories. Through the analysis of the results, it is found that when the probability of a sample belonging to a positive class is equivalent to the probability of a sample belonging to a negative class, there is a high possibility of misclassification of the sample. To address this issue, this study proposes a new SVM ensemble model construction method. We generate SVM sub-classifiers to extract error prone set, and utilise three different combination rules to construct ensemble model based on the output probability of the samples. The accuracy of the whole data set is improved for the classification accuracy of the easily misclassified data are improved. This research evaluated the model based on Wind turbine SCADA data and UCI benchmark datasets. According to the research results, the accuracy rate of SCADA data is as high as 97.52%, and the accuracy rate of German credit data is 77.50%. Meanwhile, the comparison with other methods also validates the effectiveness of the proposed method.

ENERGY Pro: Spatially explicit agent-based model on achieving positive energy districts

This article describes the ENERGY Pro agent-based model using the Overview, Design Concept, and Details + Human Decision-making (ODD+D protocol). The model is empirically explicit and aims to investigate the adoption decisions of homeowners in Amsterdam on different energy-efficient retrofitting (EER) measures. Following the ODD+D protocol, this study uncovers the conceptual framework used for model construction, the spatial microsimulation process of expanding the data, and the model implementation details. The article also describes sensitivity analysis, validation results, and how to use and adapt the model. With this article, the authors aim to make the model replicable and accessible to other researchers and inspire them using the combination of social simulation and spatial microsimulation in studying the energy transition.•The agent-based model is described using the ODD+D protocol.•The combination of simulation methods is used for constructing an empirical model.•The model on energy transition can be adapted for other cities.

Construction and validation of nomogram prognostic model for predicting survival in hepatoblastoma patients: a population-based study.

For patients with hepatoblastoma (HB), current staging system is not accurate in predicting survival outcomes. The aim of this study was to develop two accurate survival prediction models to guide clinical decision making. A retrospective analysis of 424 HB patients was performed from 2004 to 2015 using the Surveillance, Epidemiology and End Results (SEER) database. Univariate and multivariate Cox regression analysis was used to screen for variables. The identified variables were used to build survival prediction model. The performance of the nomogram models was assessed based on the concordance index (C-index), calibration plot, and receiver operating characteristic (ROC) curve. The Cox regression analysis identified six variables affecting overall survival (OS) in HB patients, including race, tumor size, lymph node involvement, distant metastases, surgery and chemotherapy. And the Cox regression analysis identified five variables including race, lymph node involvement, distant metastases, surgery, and chemotherapy that affect cancer-specific survival (CCS) in HB patients. In the training cohort, the C-index of the nomogram in predicting the OS was 0.791 [95% confidence intervals (95% CI) 0.717-0.865], CSS was 0.805(95% CI 0.728-0.882). In the validation cohort, the C-index of the nomogram in predicting the OS was 0.712 (95% CI 0.511-0.913), the CSS was 0.751 (95% CI 0.566-0.936). In the training cohort, the area under the receiver operator characteristics curve (AUC) values of the nomogram in prediction of the 1-, 3-, and 5-year OS were 0.842 (95% CI 0.739-0.944), 0.759 (95% CI 0.670-0.849), and 0.770 (95% CI 0.686-0.852), respectively. In the validation cohort, the AUC values for prediction of the 1-, 3-, and 5-year OS were 0.920 (95% CI 0.806-1.034), 0.863 (95% CI 0.750-0.976), and 0.844 (95% CI 0.721-0.967), respectively. Two nomogram models were developed and validated in this study which provided accurate prediction of the OS and CSS in HB patients. The constructed models can be used for predicting survival outcomes and guide treatment for HB patients.

MPHGCL-DDI: Meta-Path-Based Heterogeneous Graph Contrastive Learning for Drug-Drug Interaction Prediction.

The combinatorial therapy with multiple drugs may lead to unexpected drug-drug interactions (DDIs) and result in adverse reactions to patients. Predicting DDI events can mitigate the potential risks of combinatorial therapy and enhance drug safety. In recent years, deep models based on heterogeneous graph representation learning have attracted widespread interest in DDI event prediction and have yielded satisfactory results, but there is still room for improvement in prediction performance. In this study, we proposed a meta-path-based heterogeneous graph contrastive learning model, MPHGCL-DDI, for DDI event prediction. The model constructs two contrastive views based on meta-paths: an average graph view and an augmented graph view. The former represents that there are connections between drugs, while the latter reveals how the drugs connect with each other. We defined three levels of data augmentation schemes in the augmented graph view and adopted a combination of three losses in the model training phase: multi-relation prediction loss, unsupervised contrastive loss and supervised contrastive loss. Furthermore, the model incorporates indirect drug information, protein-protein interactions (PPIs), to reveal latent relations of drugs. We evaluated MPHGCL-DDI on three different tasks of two datasets. Experimental results demonstrate that MPHGCL-DDI surpasses several state-of-the-art methods in performance.

Development of a biomarker signature associated with anoikis to predict prognosis and immunotherapy response in melanoma.

Skin cutaneous melanoma (SKCM) is malignant cancer known for its high aggressiveness and unfavorable prognosis, particularly in advanced tumors. Anoikis is a specific pattern of programmed cell death associated with tumor regeneration, migration, and metastasis. Nevertheless, limited research has been conducted to investigate the function of anoikis in SKCM. Anoikis-related genes (ARGs) were extracted from Genecards to identify SKCM subtypes and to explore the immune microenvironment between the different subtypes. Prognostic models of SKCM were developed by LASSO COX regression analysis. Subsequently, the predictive value of risk scores in SKCM and the association with immunotherapy were further explored. Finally, the expression of 6 ARGs involved in the model construction was detected by immunohistochemistry and PCR. This study identified 20 ARGs significantly associated with SKCM prognosis and performed disease subtype analysis of samples based on these genes, different subtypes exhibited significantly different clinical features and tumor immune microenvironment (TIME) landscapes. The risk score prognostic model was generated by further screening and identification of the six ARGs. The model exhibited a high degree of sensitivity and specificity to predict the prognosis of individuals with SKCM. These high- and low-risk populations showed different immune statuses and drug sensitivity. Further immunohistochemical and PCR experiments identified significant differential expression of the six ARGs in tumor and normal samples. Anoikis-based features may serve as novel prognostic biomarkers for SKCM and may provide important new insights for survival prediction and individualized treatment development.

Research on Promoting Entrepreneurship among Returning Migrant Workers

To encourage migrant workers to return to their hometowns for entrepreneurship, the Guangxi government has formulated a series of supportive policies under the guidance of national policies. Understanding the direction for policy improvement requires exploring what migrant workers truly want and what concerns they have. This paper first establishes an index system based on eight dimensions including subjective attitudes, physiological motivations, and social motivations through the construction of a Random Forest model, analyzing the importance of various factors affecting the willingness of migrant workers to return home for entrepreneurship. The study found that the factor "influence from peer groups or media reports" has a significant impact on their willingness. Secondly, using the AMIRA model to predict the per capita disposable income of rural residents in Guangxi for the next seven years, it was found that their income is on an upward trend, which is conducive to enhancing the willingness of migrant workers to return home for entrepreneurship. Lastly, based on the research, countermeasures and suggestions are proposed from the perspectives of the government, market, and society to promote entrepreneurship among returning migrant workers.

Construction and alidation of a severity prediction model for acute pancreatitis based on CT severity index: A retrospective case-control study.

To construct and internally and externally validate a nomogram model for predicting the severity of acute pancreatitis (AP) based on the CT severity index (CTSI).A retrospective analysis of clinical data from 200 AP patients diagnosed at the Hefei Third Clinical College of Anhui Medical University from June 2019 to June 2022 was conducted. Patients were classified into non-severe acute pancreatitis (NSAP, n = 135) and severe acute pancreatitis (SAP, n = 65) based on final clinical diagnosis. Differences in CTSI, general clinical features, and laboratory indicators between the two groups were compared. The LASSO regression model was used to select variables that might affect the severity of AP, and these variables were analyzed using multivariate logistic regression. A nomogram model was constructed using R software, and its AUC value was calculated. The accuracy and practicality of the model were evaluated using calibration curves, Hosmer-Lemeshow test, and decision curve analysis (DCA), with internal validation performed using the bootstrap method. Finally, 60 AP patients treated in the same hospital from July 2022 to December 2023 were selected for external validation.LASSO regression identified CTSI, BUN, D-D, NLR, and Ascites as five predictive factors. Unconditional binary logistic regression analysis showed that CTSI (OR = 2.141, 95%CI:1.369-3.504), BUN (OR = 1.378, 95%CI:1.026-1.959), NLR (OR = 1.370, 95%CI:1.016-1.906), D-D (OR = 1.500, 95%CI:1.112-2.110), and Ascites (OR = 5.517, 95%CI:1.217-2.993) were independent factors influencing SAP. The established prediction model had a C-index of 0.962, indicating high accuracy. Calibration curves demonstrated good consistency between predicted survival rates and actual survival rates. The C-indexes for internal and external validation were 0.935 and 0.901, respectively, with calibration curves close to the ideal line.The model based on CTSI and clinical indicators can effectively predict the severity of AP, providing a scientific basis for clinical decision-making by physicians.

A heterogeneous graph-based semi-supervised learning framework for access control decision-making

For modern information systems, robust access control mechanisms are vital in safeguarding data integrity and ensuring the entire system’s security. This paper proposes a novel semi-supervised learning framework that leverages heterogeneous graph neural network-based embedding to encapsulate both the intricate relationships within the organizational structure and interactions between users and resources. Unlike existing methods focusing solely on individual user and resource attributes, our approach embeds organizational and operational interrelationships into the hidden layer node embeddings. These embeddings are learned from a self-supervised link prediction task based on a constructed access control heterogeneous graph via a heterogeneous graph neural network. Subsequently, the learned node embeddings, along with the original node features, serve as inputs for a supervised access control decision-making task, facilitating the construction of a machine-learning access control model. Experimental results on the open-sourced Amazon access control dataset demonstrate that our proposed framework outperforms models using original or manually extracted graph-based features from previous works. The prepossessed data and codes are available on GitHub,facilitating reproducibility and further research endeavors.

Construction and application of economic resilience evaluation model for megacities.

Economic resilience provides a new perspective for megacities to achieve sustainable development when facing multiple shocks, and its accurate evaluation is an essential prerequisite for optimizing urban governance. There are currently no generally accepted methods for empirical evaluation or measuring economic resilience, and the present study aims to contribute to in both the research field and methodology. The present study sets dimensions and indicators based on economic resilience's theoretical and empirical research and used Decision Making Trial and Evaluation Laboratory (DEMATEL) and Interactive Structural Modeling (ISM) methods to exclude the effect indicators and divide the indicator hierarchy, respectively. Subsequently, the present study conducts model validation using Chinese megacities as a case study. The game theory weighting method, which combines the Analytic Hierarchy Process (AHP) and Entropy methods, is used to calculate indicator weights, and the VIKOR (VIseKriterijumska Optimizacija i KOmpromisno Resenje) method is used to evaluate and compare economic resilience of megacities. The research findings indicate that the evaluation model constructed in the present study included 15 indicators (after excluding three effect indicators) divided into four levels. After merging the levels, they correspond to three dimensions: resistance, recoverability, and adaptability. In addition, using Chinese megacities as a case study, the evaluation results found that Beijing, Shanghai, and Shenzhen have high economic resilience, Tianjin and Guangzhou have moderate economic resilience, Chengdu has low economic resilience, and Chongqing has the lowest economic resilience. This result is consistent with previous studies and verifies the model's effectiveness. The present study also found that megacities with lower levels of economic resilience exhibit a more significant upward trend, as well as the highest and higher proportion of economic resilience in Chinese megacities depending on time passes, indicating that megacities' economic resilience is weakening. The evaluation result obtained in the present study is more specific, precise, and focused on depicting the distribution differences and development trends of economic resilience at the urban level.

Analysis of employee diligence and mining of behavioral patterns based on portrait portrayal

With the deepening of enterprise digital construction, the portrait portrayal based on employee behaviors has gradually become a research focus. Currently, the employee's portrait portrayal mostly has the problems of simple means, low efficiency, limited solving ability, etc., making the results more one-sided. Therefore, a data mining-based employee portrait portrayal model is proposed. The content of employee portrait portrayal is deeply analyzed, and the overall framework of the model is designed. A diligence analysis model (DAM) based on improved GAN is constructed, and the diligence evaluation of employees is clarified to realize the diligence evaluation. The results of diligence analysis of DAM have high accuracy (80.39%) and outperform SA (70.24%), K-means (51.79%) and GAN (67.25%). The Kappa coefficient of DAM reaches 0.7384, which is highly consistent and higher than SA (0.6075), K-means (0.3711) and GAN (0.5661). The Local Outlier Factor (LOF) and Isolation Forest (IF) are used to detect abnormal behaviors on the employees, and mine the abnormal behavior patterns on different granularity time. The LSTM model (Att-LSTM) based on the attention mechanism is used to complete the prediction of employees' software usage behaviors, and analyze and summarize the characteristics of employee's behaviors from multiple perspectives. Att-LSTM predicts the best with an RMSE of 0.82983, which is better than LSTM (0.90833) and SA (0.97767); AM-LSTM has a MAPE of 0.80323, which is better than LSTM (0.86233) and SA (0.92223). The results show that the data mining-based employee portrait portrayal method can better solve the problem of enterprise employees' digital construction, and provide a new way of thinking for the construction of enterprise-level employees' digital portrait model and the analysis of employee behavior.

Preferences for and intention to use an app for premenstrual mental health symptoms using the Health Behaviour Model (HBM)

Premenstrual symptoms are common, with premenstrual syndrome and premenstrual dysphoric disorder associated with decreased wellbeing and increased suicidality. Apps can offer convenient support for premenstrual mental health symptoms. We aimed to understand app preferences and Health Belief Model (HBM) constructs driving app use intention. An online survey was delivered. Structural equation modelling (SEM) explored HBM constructs. Data from 530 United Kingdom based participants who reported their mental health was impacted by their menstrual cycle (mean age = 35.85, SD = 7.28) were analysed. In terms of preferred app features, results indicated that symptom monitoring (74.72%, n = 396) and psychoeducation (57.92%, n = 307) were sought after, with 52.64% (n = 279) indicating unwillingness to pay for an app for mental health symptoms related to the menstrual cycle. Regarding HBM results, Satorra–Bentler-scaled fit statistics indicated a good model fit (χ2(254) = 565.91, p < 0.001; CFI = 0.939, RMSEA = 0.048, SRMR = 0.058). HBM constructs explained 58.22% of intention to use, driven by cues to action (β = 0.49, p < 0.001), perceived barriers (β = −0.22, p < 0.001), perceived severity (β = 0.16, P = 0.012), and perceived benefits (β = 0.10, p = 0.035). Results indicate that app developers should undertake co-design, secure healthcare professional endorsement, highlight therapeutic benefits, and address barriers like digital discomfort, privacy concerns, and quality.

Predicting leachate impact on groundwater using electrical conductivity and oxidation–reduction potential measurements: An empirical and theoretical approach

This study developed innovative predictive models of groundwater pollution using in situ electrical conductivity (EC) and oxidation–reduction potential (ORP) measurements at livestock carcass burial sites. Combined electrode analysis (EC and ORP) and machine learning techniques efficiently and accurately distinguished between leachate and background groundwater. Two models—empirical and theoretical—were constructed based on a supervised classification framework. The empirical model constructs a classifier with high accuracy, sensitivity, and specificity, utilizing the comprehensive in situ EC and ORP measurements. The theoretical model with only two end members achieves comparable performance by simulating the leachate–groundwater interactions using a geochemical mixing model. Besides enhancing the early detection capabilities, our approach considerably reduces the reliance on extensive hydrochemical analyses, thus streamlining the monitoring process. Moreover, the use of field parameters was found to proactively identify potential pollution incidents, enhancing the efficiency of groundwater monitoring strategies. Our approach is applicable to various waste disposal sites, indicating its extensive potential for environmental monitoring and management.

CMIP6 ESMs overestimate greening and the photosynthesis trends in Dryland East Asia

The Dryland East Asia (DEA) is one of the largest inland arid regions, and vegetation is very sensitive to climate change. The complex environment in DEA with defects of modeling construction make it difficult to simulate and predict changes in vegetation structure and productivity. Here, we use the emergent constraint (EC) method to constrain the future interannual leaf area index (LAI) and gross primary productivity (GPP) trends in DEA, under four scenarios of the latest Sixth Coupled Model Intercomparison Project (CMIP6) model ensemble. LAI and GPP increase in all scenarios in the near term (2015–2050), with continued growth in SSP370 and SSP585 and stasis in SSP126 and SSP245 in the far term (2051–2100). However, after building effective EC relationships, the constrained increasing trends of LAI (GPP) are reduced by 43.5 %–53.9 % (30.5 %–50.0 %) compared with the uncertainties of the original ensemble, which are reduced by 10.0 %–45.7 % (4.6 %–34.3 %). We also extend the EC in moving windows and grid cells, further strengthening the robustness of the constraints, especially by illustrating spatial sources of these emergent relationships. Overestimations of LAI and GPP trends suggest that current CMIP6 models may be insufficient to capture the complex relationships between climate change and vegetation dynamics in DEA; however, these models can be adjusted based on established emergent relationships.

Physiologically-based pharmacokinetic modeling to predict the exposure and provide dosage regimens of Ustekinumab in pediatric patients with inflammatory bowel disease

Ustekinumab (UST), a fully human immunoglobulin G1 κ monoclonal antibody, exhibiting high affinity for the p40 subunit shared by IL-12 and IL-23, which play key roles in the pathogenesis of inflammatory bowel disease (IBD). By scaling the physiologically-based pharmacokinetic modeling (PBPK) model of UST in adult patients with IBD, we aim to predict effective dosages for UST in pediatric patients, thereby offering a more practical dosing regimen for real-world applications. In this work, a PBPK model for UST in adult patients with IBD has been developed using PK-Sim and Mobi. Advanced ontogeny model has been incorporated to extrapolate the model to pediatric patients. The simulation results showed that the fold errors of the predicted and observed values of the area under the curve (AUC) and peak plasma concentration (Cmax) were between 0.79 and 1.73. For children aged 6–18, it is recommended to administer the drug per kilogram of body weight, at the model-recommended dose, to achieve a median AUC similar to that of the adult reference population post-administration. This comprehensive model construction enables us to comprehensively and extensively explore the pharmacokinetic characteristics of UST in pediatric patients of different age groups, providing robust support for clinical applications and personalized drug therapy.

Development and validation of a machine learning-based prediction model for sudden cardiac death in the general population: insights from the community cardiovascular cohort

Abstract Background/Introduction Accurately predicting and managing the risk of sudden cardiac death (SCD) in the general population is a formidable challenge. Purpose This study aimed to develop and validate a machine learning (ML)-based prediction model for SCD using data from the Community Cardiovascular Cohort (CCCC) in Taiwan. Methods A community-based cohort study was used for model construction . Four ML algorithms, including extreme gradient boosting (XGBoost), random forests (RF), logistic regression, and deep neural network (DNN), underwent meticulous evaluation. Performance assessment encompassed receiver operating characteristic (ROC) curves, area under the curves (AUC), and accuracy metrics. Significant risk factors were identified through feature importance analysis. Results We employed a random partitioning strategy, allocating 1744 participants (55%) to the training dataset, 952 participants (30%) to the test dataset, and 476 participants (15%) to the validation dataset. In the data training stage, the XGBoost algorithm demonstrated outstanding prediction performance for SCD (AUC: 1.000); In the validation stage, the XGBoost algorithm demonstrated good prediction performance for SCD prediction (AUC: 0.791, 95%CI: 0.617-0.964; Accuracy: 0.971). Using XGBoost-generated predicted probabilities (P(XGB)), we identified individuals with P(XGB) ≥ 0.5 (Group 1) who displayed significantly higher SCD risk (HR 32.0; 95% CI: 9.29-109.9) compared to those with P(XGB) &amp;lt; 0.5 (Group 0). Conclusions Our study demonstrates the effectiveness of ML algorithms, particularly XGBoost, in predicting event of SCD in the general population. The identification of these pivotal risk factors holds profound implications, empowering the identification of high-risk individuals and facilitating the implementation of precisely tailored preventive strategies.Figure 1:KM sruvival curveFigure 2:Feature importance analysis

Hyperspectral Image Classification Based on Two-Branch Multiscale Spatial Spectral Feature Fusion with Self-Attention Mechanisms

In recent years, the use of deep neural network in effective network feature extraction and the design of efficient and high-precision hyperspectral image classification algorithms has gradually become a research hotspot for scholars. However, due to the difficulty of obtaining hyperspectral images and the high cost of annotation, the training samples are very limited. In order to cope with the small sample problem, researchers often deepen the network model and use the attention mechanism to extract features; however, as the network model continues to deepen, the gradient disappears, the feature extraction ability is insufficient, and the computational cost is high. Therefore, how to make full use of the spectral and spatial information in limited samples has gradually become a difficult problem. In order to cope with such problems, this paper proposes two-branch multiscale spatial–spectral feature aggregation with a self-attention mechanism for a hyperspectral image classification model (FHDANet); the model constructs a dense two-branch pyramid structure, which can achieve the high efficiency extraction of joint spatial–spectral feature information and spectral feature information, reduce feature loss to a large extent, and strengthen the model’s ability to extract contextual information. A channel–space attention module, ECBAM, is proposed, which greatly improves the extraction ability of the model for salient features, and a spatial information extraction module based on the deep feature fusion strategy HLDFF is proposed, which fully strengthens feature reusability and mitigates the feature loss problem brought about by the deepening of the model. Compared with five hyperspectral image classification algorithms, SVM, SSRN, A2S2K-ResNet, HyBridSN, SSDGL, RSSGL and LANet, this method significantly improves the classification performance on four representative datasets. Experiments have demonstrated that FHDANet can better extract and utilise the spatial and spectral information in hyperspectral images with excellent classification performance under small sample conditions.