Search Results (1,679)

This paper introduces novel inverse optimization algorithms (RC and DC) for neural network training in stock price forecasting in an attempt to overcome the traditional gradient descent limitation of local minima convergence. The key novelty is a stochastic algorithm for inverse problems adapted to neural network training, where target function values decrease iteratively through selective weight modification. Experimental analysis used closing price data from 40 Russian companies, comparing traditional activation functions (linear, sigmoid, tanh) with specialized functions (sincos, cloglogm, mish) across perceptrons and single-hidden-layer networks. Key findings show the superiority of the DC method for single-layer networks, while RC proves most effective for hidden-layer networks. The linear activation function with the RC algorithm delivered optimal results in most experiments, challenging conventional nonlinear activation preferences. The optimal architecture, namely, a single hidden layer with two neurons, achieved the best prediction accuracy in 70% of cases. The research confirms that inverse optimization algorithms can provide higher training efficiency than classical gradient methods, offering practical improvements for financial forecasting. Full article

Objective: This study investigated associations between physical activity (PA) during late adolescence and emerging adulthood and bone strength in emerging adulthood by utilizing advanced finite element analysis of computed tomography (CT/FEA) technology beyond the traditional dual-energy X-ray absorptiometry (DXA) method. Methods: This study included 266 participants (152 females) from the Iowa Bone Development Study. PA volume (average acceleration) and intensity (intensity gradient) metrics were calculated from ActiGraph accelerometer data collected at ages 17, 19, 21, and 23 years. Compressive modulus and compressive stiffness of the tibia were estimated at age 23 via CT/FEA of the tibia. Sex-specific linear regression models were used to evaluate associations between PA metrics and bone outcomes, adjusting for age, height, weight, musculoskeletal fitness, and calcium intake. Results: Intensity gradient averaged over 17–23 years of age was positively associated with compressive stiffness at age 23 years in both females and males (p < 0.01). Intensity gradient was positively associated with compressive modulus in females (p < 0.01), but not in males. No significant associations were found between average acceleration and either compressive stiffness or modulus in either sex (p > 0.05). Conclusions: Using a state-of-the-art CT/FEA method, this study suggests that high-intensity PA during late adolescence and emerging adulthood improves bone strength. Full article

This study presents a unified machine learning strategy for identifying various degrees of sarcopenia severity in older adults. The approach combines three optimized algorithms (Random Forest, Gradient Boosting, and Multilayer Perceptron) into a stacked ensemble model, which is assessed with clinical data. A thorough data preparation process involved synthetic minority oversampling to ensure class balance and a dual approach to feature selection using Least Absolute Shrinkage and Selection Operator regression and Random Forest importance. The integrated model achieved remarkable performance with an accuracy of 96.99%, an F1 score of 0.9449, and a Cohen’s Kappa coefficient of 0.9738 while also demonstrating excellent calibration (Brier Score: 0.0125). Interpretability analysis through SHapley Additive exPlanations values identified appendicular skeletal muscle mass, body weight, and functional performance metrics as the most significant predictors, enhancing clinical relevance. The ensemble approach showed superior generalization across all sarcopenia classes compared to individual models. Although limited by dataset representativeness and the use of conventional multiclass classification techniques, the framework shows considerable promise for non-invasive sarcopenia risk assessments and exemplifies the value of interpretable artificial intelligence in geriatric healthcare. Full article

Diabetic Retinopathy (DR) is a leading cause of vision impairment globally, underscoring the need for accurate and early diagnosis to prevent disease progression. Although fundus imaging serves as a cornerstone of Computer-Aided Diagnosis (CAD) systems, several challenges persist, including lesion scale variability, blurry morphological patterns, inter-class imbalance, limited labeled datasets, and computational inefficiencies. To address these issues, this study proposes an end-to-end diagnostic framework that integrates an enhanced preprocessing pipeline with a novel deep learning architecture, Hierarchical-Inception-Residual-Dense Network (HIRD-Net). The preprocessing stage combines Contrast Limited Adaptive Histogram Equalization (CLAHE) with Dilated Difference of Gaussian (D-DoG) filtering to improve image contrast and highlight fine-grained retinal structures. HIRD-Net features a hierarchical feature fusion stem alongside multiscale, multilevel inception-residual-dense blocks for robust representation learning. The Squeeze-and-Excitation Channel Attention (SECA) is introduced before each Global Average Pooling (GAP) layer to refine the Feature Maps (FMs). It further incorporates four GAP layers for multi-scale semantic aggregation, employs the Hard-Swish activation to enhance gradient flow, and utilizes the Focal Loss function to mitigate class imbalance issues. Experimental results on the IDRiD-APTOS2019, DDR, and EyePACS datasets demonstrate that the proposed framework achieves 93.46%, 82.45% and 79.94% overall classification accuracy using only 4.8 million parameters, highlighting its strong generalization capability and computational efficiency. Furthermore, to ensure transparent predictions, an Explainable AI (XAI) approach known as Gradient-weighted Class Activation Mapping (Grad-CAM) is employed to visualize HIRD-Net’s decision-making process. Full article

In this study, the key challenges faced by global navigation satellite systems (GNSSs) in the field of security are addressed, and an eye diagram-based deep learning framework for intelligent classification of interference types is proposed. GNSS signals are first transformed into two-dimensional eye diagrams, enabling a novel visual representation wherein interference types are distinguished through entropy-centric feature analysis. Specifically, the quantification of information entropy within these diagrams serves as a theoretical foundation for extracting salient discriminative features, reflecting the structural complexity and uncertainty of the underlying signal distortions. We designed a hybrid architecture that integrates spatial feature extraction, gradient stability enhancement, and time dynamics modeling capabilities and combines the advantages of a convolutional neural network, residual network, and bidirectional long short-term memory network. To further improve model performance, we propose an improved coati optimization algorithm (ICOA), which combines chaotic mapping, an elite perturbation mechanism, and an adaptive weighting strategy for hyperparameter optimization. Compared with mainstream optimization methods, this algorithm improves the convergence accuracy by more than 30%. Experimental results on jamming datasets (continuous wave interference, chirp interference, pulse interference, frequency-modulated interference, amplitude-modulated interference, and spoofing interference) demonstrate that our method achieved performance in terms of accuracy, precision, recall, F1 score, and specificity, with values of 98.02%, 97.09%, 97.24%, 97.14%, and 99.65%, respectively, which represent improvements of 1.98%, 2.80%, 6.10%, 4.59%, and 0.33% over the next-best model. This study provides an efficient, entropy-aware, intelligent, and practically feasible solution for GNSS interference identification. Full article

Fish diseases are one of the primary causes of economic losses in aquaculture. Existing deep learning models have progressed in fish disease detection and lesion segmentation. However, many models still have limitations, such as detecting only a single type of fish disease or completing only a single task within fish disease detection. To address these limitations, we propose FDMNet, a multi-task learning network. Built upon the YOLOv8 framework, the network incorporates a semantic segmentation branch with a multi-scale perception mechanism. FDMNet performs detection and segmentation simultaneously. The detection and segmentation branches use the C2DF dynamic feature fusion module to address information loss during local feature fusion across scales. Additionally, we use uncertainty-based loss weighting together with PCGrad to mitigate conflicting gradients between tasks, improving the stability and overall performance of FDMNet. On a self-built image dataset containing three common fish diseases, FDMNet achieved 97.0% mAP50 for the detection task and 85.7% mIoU for the segmentation task. Relative to the multi-task YOLO-FD baseline, FDMNet’s detection mAP50 improved by 2.5% and its segmentation mIoU by 5.4%. On the dataset constructed in this study, FDMNet achieved competitive accuracy in both detection and segmentation. These results suggest potential practical utility. Full article

This study leverages multi-source remote sensing data—Nighttime Light (NTL) imagery and POI (Point of Interest) datasets—to quantify the spatiotemporal interaction between urban spatial restructuring and logistics industry evolution in Zhengzhou, China. Using calibrated NPP/VIIRS NTL data (2012–2022) and fine-grained POI data, we (1) identified urban functional spaces through kernel density-based spatial grids weighted by public awareness parameters; (2) extracted built-up areas via the dynamic adaptive threshold segmentation of NTL gradients; (3) analyzed logistics agglomeration dynamics using emerging spatiotemporal hotspot analysis (ESTH) and space–time cube models. The results show that Zhengzhou’s urban form transitioned from a monocentric to a polycentric structure, with NTL trajectories revealing logistics hotspots expanding along air–rail multimodal corridors. POI-derived functional spaces shifted from single-dominant to composite patterns, while ESTH detected policy-driven clusters in Airport Economic Zones and market-driven suburban cold chain hubs. Bivariate LISA confirmed the spatial synergy between logistics growth and urban expansion, validating the “policy–space–industry” interaction framework. This research demonstrates how integrated NTL-POI remote sensing techniques can monitor policy impacts on urban systems, providing a replicable methodology for sustainable logistics planning. Full article

Lithology identification is a critical technology for geological resource exploration and engineering safety assessment. However, traditional methods suffer from insufficient feature representation and low classification accuracy due to challenges such as weathering, vegetation cover, and spectral overlap in complex sedimentary rock regions. This study proposes a hierarchical multi-feature random forest algorithm based on Feature-Preserved Compressive Sampling (FPCS). Using 3D laser point cloud data from the Manas River outcrop in the southern margin of the Junggar Basin as the test area, we integrate graph signal processing and multi-scale feature fusion to construct a high-precision lithology identification model. The FPCS method establishes a geologically adaptive graph model constrained by geodesic distance and gradient-sensitive weighting, employing a three-tier graph filter bank (low-pass, band-pass, and high-pass) to extract macroscopic morphology, interface gradients, and microscopic fracture features of rock layers. A dynamic gated fusion mechanism optimizes multi-level feature weights, significantly improving identification accuracy in lithological transition zones. Experimental results on five million test samples demonstrate an overall accuracy (OA) of 95.6% and a mean accuracy (mAcc) of 94.3%, representing improvements of 36.1% and 20.5%, respectively, over the PointNet model. These findings confirm the robust engineering applicability of the FPCS-based hierarchical multi-feature approach for point cloud lithology identification. Full article

Spatial heterogeneity in soil deposition poses a significant challenge to accurate geotechnical characterization, which is essential for sustainable infrastructure development. This study presents an advanced geotechnical data-driven mapping framework, based on a monotonized and augmented formulation of Shepard’s inverse distance weighting (IDW) algorithm, implemented through the Google Earth Engine (GEE) platform. The approach is rigorously evaluated through a comparative analysis against the classical IDW and Kriging techniques using standard key performance indices (KPIs). Comprehensive field and laboratory data repositories were developed in accordance with international geotechnical standards (e.g., ASTM). Key geotechnical parameters, i.e., standard penetration test (SPT-N) values, shear wave velocity (Vs), soil classification, and plasticity index (PI), were used to generate high-resolution geospatial models for a previously unmapped region, thereby providing essential baseline data for building infrastructure design. The results indicate that the augmented IDW approach exhibits the best spatial gradient conservation and local anomaly detection performance, in alignment with Tobler’s First Law of Geography, and outperforms Kriging and classical IDW in terms of predictive accuracy and geologic plausibility. Compared to classical IDW and Kriging, the augmented IDW algorithm achieved up to a 44% average reduction in the RMSE and MAE, along with an approximately 30% improvement in NSE and PC. The difference in spatial areal coverage was found to be up to 20%, demonstrating an improved capacity to model spatial subsurface heterogeneity. Thematic design maps of the load intensity (LI), safe bearing capacity (SBC), and optimum foundation depth (OD) were constructed for ready application in practical design. This work not only establishes the inadequacy of conventional geostatistical methods in highly heterogeneous soil environments but also provides a scalable framework for geotechnical mapping with accuracy in data-poor environments. Full article

Streak Tube Imaging Lidar (STIL) offers significant advantages in long-range sensing and ultrafast diagnostics by encoding spatial-temporal information as streaks, and hence decodes 3D images using tailored algorithm. However, under low-photon conditions that caused either long-range or reduced exposure time, the reconstructed image suffer from low contrast, strong noise and blurring, hindering the application in various scenarios. To address this challenge, we propose a Multi-Sparsity Constraints and Adaptive Regularization (MSC-AR) algorithm based on the Maximum a Posteriori (MAP) framework, which jointly denoises and deblurs degraded streak images and efficiently solved using the Alternating Direction Method of Multipliers (ADMM). MSC-AR considers gradient sparsity, intensity sparsity, and an adaptively weighted Total Variation (TV) regularization along the temporal dimension of the streak image which collaboratively optimizing image quality and structural detail, thus better 3D restoration results in low-photon conditions. Experimental results demonstrate that MSC-AR significantly outperforms existing approaches under low-photon conditions. At an exposure time of 300 ms, it achieves millimeter-level RMSE and over 88% SSIM in depth image reconstruction, while maintaining robustness and generalization across different reconstruction strategies and target types. Full article

The rapid development of wireless communication technology is leading to increasingly scarce spectrum resources, making efficient utilization a critical challenge. This paper proposes a Convolutional Neural Network–Long Short-Term Memory-Integrated Gradient-Weighted Class Activation Mapping (GC-CNN-LSTM) model, aimed at enhancing the accuracy of long-term spectrum prediction across multiple frequency bands and improving model interpretability. First, we achieve multi-frequency long-term spectrum prediction using a CNN-LSTM and compare its performance against models including LSTM, GRU, CNN, Transformer, and CNN-LSTM-Attention. Next, we use an improved Grad-CAM method to explain the model and obtain global heatmaps in the time–frequency domain. Finally, based on these interpretable results, we optimize the input data by selecting high-importance frequency points and removing low-importance time segments, thereby enhancing prediction accuracy. The simulation results show that the Grad-CAM-based approach achieves good interpretability, reducing RMSE and MAPE by 6.22% and 4.25%, respectively, compared to CNN-LSTM, while a similar optimization using SHapley Additive exPlanations (SHAP) achieves reductions of 0.86% and 3.55%. Full article

Flash floods arise from the interaction of rugged topography, short-duration intense rainfall, and rapid flow concentration. Conventional risk mapping often builds empirical indices with expert-assigned weights or trains supervised models on historical event inventories—approaches that degrade in data-scarce regions. We propose a fully data-driven, unsupervised Geographic Information System (GIS) framework based on fractional order k-means, which clusters multi-dimensional geospatial features without labeled flood records. Five raster layers—elevation, slope, aspect, 24 h maximum rainfall, and distance to the nearest stream—are normalized into a feature vector for each 30 m × 30 m grid cell. In a province-scale case study of Zhejiang, China, the resulting risk map aligns strongly with the observations: 95% of 1643 documented flash flood sites over the past 60 years fall within the combined high- and medium-risk zones, and 65% lie inside the high-risk class. These outcomes indicate that the fractional order distance metric captures physically realistic hazard gradients while remaining label-free. Because the workflow uses commonly available GIS inputs and open-source tooling, it is computationally efficient, reproducible, and readily transferable to other mountainous, data-poor settings. Beyond reducing subjective weighting inherent in index methods and the data demands of supervised learning, the framework offers a pragmatic baseline for regional planning and early-stage screening. Full article

Lubricating oil reflects mechanical component aging and wear. Accurate quantification of its wear metals is essential for equipment safety and intelligent maintenance. This study introduces a rapid, non-destructive method for detecting wear metal content in lubricating oil using hyperspectral technology to overcome limitations such as bulky, expensive instruments and destructive testing in current spectroscopic techniques. Absorption spectra of 98 marine gearbox oil samples were acquired using Hach UV-Vis and GLT optical fiber spectrometers. We propose a multi-head attention mechanism enhanced genetic algorithm (MHA-GA) for deep feature extraction, integrating attention weights into band selection and fitness evaluation to identify key features under multi-element interference. Wear metal prediction models were constructed using random forest (RF), support vector regression (SVR), and extreme gradient boosting (XGBoost). Results demonstrate MHA-GA outperformed traditional genetic algorithm (GA) and competitive adaptive reweighted sampling (CARS) in feature selection. The MHA-GA-XGBoost model performed best. Fe prediction R² reached 0.96 (Hach) and 0.93 (GLT), with RPDs of 5.33 and 3.90. For Cu, R² reached 0.91 and 0.83, with RPDs of 3.35 and 2.42. The results indicate that hyperspectral technology combined with machine learning enables effective non-destructive wear metal quantification, offering a promising strategy for intelligent maintenance and condition monitoring of lubricating oil. Full article

This study presents a gradient-informed proxy initialization framework designed to improve training efficiency and predictive performance in deep learning models for time-series forecasting. The method extends the Laor Initialization approach by introducing backward gradient norm clustering as a selection criterion for input-layer weights, evaluated through a lightweight, architecture-agnostic proxy model. Only the numerical input layer adopts the selected initialization, while internal components retain standard schemes such as Xavier, Kaiming, or Orthogonal, maintaining compatibility and reducing overhead. The framework is evaluated on a real-world financial forecasting task: identifying high-risk equities from the Thai Market Surveillance Measure List, a domain characterized by label imbalance, non-stationarity, and limited data volume. Experiments across five architectures, including Transformer, ConvTran, and MMAGRU-FCN, show that the proposed strategy improves convergence speed and classification accuracy, particularly in deeper and hybrid models. Results in recurrent-based models are competitive but less pronounced. These findings support the method’s practical utility and generalizability for forecasting tasks under real-world constraints. Full article

A diverse array of soil microorganisms exhibit plant growth-promoting (PGP) traits, many of which enhance root growth and development. These microorganisms include various taxa of bacteria, fungi, microalgae and yeasts—some of which are currently used in biofertilizers and biostimulant formulations. Recent studies have begun to unravel the complex communication between plant roots and beneficial microorganisms, revealing mechanisms that modulate root nitrogen (N) uptake beyond atmospheric N₂ fixation pathways. Root N uptake is tightly regulated by plants through multiple mechanisms. These include transcriptional and post-transcriptional control of plasma membrane-localized N transporters in the epidermis, endodermis, and xylem parenchyma. Additionally, N uptake efficiency is influenced by vacuolar N storage, assimilation of inorganic N into organic compounds, and the maintenance of electrochemical gradients across root cell membranes. Many of these processes are modulated by microbial signals. This review synthesizes current knowledge on how soil microorganisms influence root N uptake, with a focus on signaling molecules released by soil beneficial microbes. These signals include phytohormones, volatile organic compounds (VOCs), and various low-molecular-weight organic compounds that affect transporter expression, root architecture, and cellular homeostasis. Special attention is paid to the molecular and physiological pathways through which these microbial signals enhance plant N acquisition and overall nutrient use efficiency. Full article