Search Results (189)

Stock price forecasting remains a persistent challenge in time series analysis due to complex inter-stock relationships and dynamic textual signals such as financial news. While Graph Neural Networks (GNNs) can model relational structures, they often struggle with capturing higher-order dependencies and are sensitive to noise. Moreover, sentiment signals are typically aggregated using fixed time windows, which may introduce temporal bias. To address these issues, we propose DASF-Net (Diffusion-Aware Sentiment Fusion Network), a multimodal framework that integrates structural and textual information for robust prediction. DASF-Net leverages diffusion processes over two complementary financial graphs—one based on industry relationships, the other on fundamental indicators—to learn richer stock representations. Simultaneously, sentiment embeddings extracted from financial news using FinBERT are aggregated over an empirically optimized window to preserve temporal relevance. These modalities are fused via a multi-head attention mechanism and passed to a temporal forecasting module. DASF-Net integrates daily stock prices and news sentiment, using a 3-day sentiment aggregation window, to forecast stock prices over daily horizons (1–3 days). Experiments on 12 large-cap S&P 500 stocks over four years demonstrate that DASF-Net outperforms competitive baselines, achieving up to 91.6% relative reduction in Mean Squared Error (MSE). Results highlight the effectiveness of combining graph diffusion and sentiment-aware features for improved financial forecasting. Full article

In view of the bottleneck problems of serious spectral peak cross-interference, redundant data dimensions, and inefficient traditional dimensionality reduction methods in the infrared spectral analysis of mixed gases, this paper studies a joint dimensionality reduction strategy combining stacked self encoder (SAE) and partial least squares (PLS) and constructs an XGBoost regression model for quantitative detection. The experimental data are from the real infrared spectrum dataset of the National Institute of Standards and Technology (NIST) database, covering key industrial gases such as CO, CH₄, etc. Compared with the traditional principal component analysis (PCA), which relies on the variance contribution rate and leads to dimensional redundancy, and the calculation efficiency of dimension parameters that need to be cross-verified for PLS dimension reduction alone, the SAE-PLS joint strategy has two advantages: first, the optimal dimension reduction is automatically determined by SAE’s nonlinear compression mechanism, which effectively overcomes the limitations of linear methods in spectral nonlinear feature extraction; and second, the feature selection is carried out by combining the variable importance projection index of PLS. Compared with SAE, the compression efficiency is significantly improved. The XGBoost model was selected because of its adaptability to high-dimensional sparse data. Its regularization term and feature importance weighting mechanism can suppress the interference of spectral noise. The experimental results show that the mean square error (MSE) on the test set is reduced to 0.012% (71.4% lower than that of random forest), and the correlation coefficient (R²) is 0.987. By integrating deep feature optimization and ensemble learning, this method provides a new solution with high efficiency and high precision for industrial process gas monitoring. Full article

The rapid growth of the Internet of Things (IoT) has revolutionized various industries by enabling interconnected devices to exchange data seamlessly. However, IoT systems face significant security challenges due to decentralized architectures, resource-constrained devices, and dynamic network environments. These challenges include denial-of-service (DoS) attacks, anomalous network behaviors, and data manipulation, which threaten the security and reliability of IoT ecosystems. New methods based on machine learning have been reported in the literature, addressing topics such as intrusion detection and prevention. This paper proposes an advanced anomaly detection framework for IoT networks expressed in several phases. In the first phase, data preprocessing is conducted using techniques like the Median-KS Test to remove noise, handle missing values, and balance datasets, ensuring a clean and structured input for subsequent phases. The second phase focuses on optimal feature selection using a Genetic Algorithm enhanced with eagle-inspired search strategies. This approach identifies the most significant features, reduces dimensionality, and enhances computational efficiency without sacrificing accuracy. In the final phase, an ensemble classifier combines the strengths of the Decision Tree, Random Forest, and XGBoost algorithms to achieve the accurate and robust detection of anomalous behaviors. This multi-step methodology ensures adaptability and scalability in handling diverse IoT scenarios. The evaluation results demonstrate the superiority of the proposed framework over existing methods. It achieves a 12.5% improvement in accuracy (98%), a 14% increase in detection rate (95%), a 9.3% reduction in false positive rate (10%), and a 10.8% decrease in false negative rate (5%). These results underscore the framework’s effectiveness, reliability, and scalability for securing real-world IoT networks against evolving cyber threats. Full article

Noise reduction for manufacturing enterprises is favorable for workers because it relieves occupational diseases and improves productivity. An acoustic metamaterial with parallel, unequal cavities is proposed and optimized, aiming to achieve an optimal broadband sound absorber in the low–frequency range with a limited total thickness. A theoretical model for the acoustic metamaterial of a hexagonal column with 6 triangular cavities and 12 right–angled trapezoidal cavities was established. The lengths of these embedded apertures were optimized using the particle swarm optimization algorithm, with initial parameters obtained from acoustic finite element simulation. Additionally, the impacts of manufacturing errors on different regions were analyzed. The experimental results prove that the proposed acoustic metamaterials can achieve an average absorption coefficient of 0.87 from 384 Hz to 667 Hz with a thickness of 50 mm, 0.83 from 265 Hz to 525 Hz with a thickness of 70 mm, and 0.82 from 156 Hz to 250 Hz with a thickness of 100 mm. The experimental validation demonstrates the accuracy of the finite element model and the effectiveness of the optimization algorithm. This extensible acoustic metamaterial, with excellent sound absorption performance in the low-frequency range, can be mass-produced and widely applied for noise control in industries. Full article

Chinese, a tonal language with inherent homophonic ambiguity, poses significant challenges for semantic disambiguation in natural language processing (NLP), hindering applications like speech recognition, dialog systems, and assistive technologies. Traditional static disambiguation methods suffer from poor adaptability in dynamic environments and low-frequency scenarios, limiting their real-world utility. To address these limitations, we propose BLAF—a novel MacBERT-BiLSTM Hybrid Architecture—that synergizes global semantic understanding with local sequential dependencies through dynamic multimodal feature fusion. This framework incorporates innovative mechanisms for the principled weighting of heterogeneous features, effective alignment of representations, and sensor-augmented cross-modal learning to enhance robustness, particularly in noisy environments. Employing a staged optimization strategy, BLAF achieves state-of-the-art performance on the SIGHAN 2015 (data fine-tuning and supplementation): 93.37% accuracy and 93.25% F1 score, surpassing pure BERT by 15.74% in accuracy. Ablation studies confirm the critical contributions of the integrated components. Furthermore, the sensor-augmented module significantly improves robustness under noise (speech SNR to 18.6 dB at 75 dB noise, 12.7% reduction in word error rates). By bridging gaps among tonal phonetics, contextual semantics, and computational efficiency, BLAF establishes a scalable paradigm for robust Chinese homophone disambiguation in industrial NLP applications. This work advances cognitive intelligence in Chinese NLP and provides a blueprint for adaptive disambiguation in resource-constrained and dynamic scenarios. Full article

In fiber Bragg grating (FBG) sensing demodulation systems, high-precision peak detection is a core requirement for demodulation algorithms. However, practical spectral signals are often susceptible to environmental noise interference, which leads to significant degradation in the accuracy of traditional demodulation methods. This study proposes a self-adaptive K-SVD (SAK-SVD) denoising algorithm based on adaptive window parameter optimization, establishing a closed-loop iterative feedback mechanism through dual iterations between dictionary learning and parameter adjustment. This approach achieves a synergistic enhancement of noise suppression and signal fidelity. First, a dictionary learning framework based on K-SVD is constructed for initial denoising, and the peak feature region is extracted by differentiating the denoised signals. By constructing statistics on the number of sign changes, an adaptive adjustment model for the window size is established. This model dynamically tunes the window parameters in dictionary learning for iterative denoising, establishing a closed-loop architecture that integrates denoising evaluation with parameter optimization. The performance of SAK-SVD is evaluated through three experimental scenarios, demonstrating that SAK-SVD overcomes the rigid parameter limitations of traditional K-SVD in FBG spectral processing, enhances denoising performance, and thereby improves wavelength demodulation accuracy. For denoising undistorted waveforms, the optimal mean absolute error (MAE) decreases to 0.300 pm, representing a 25% reduction compared to the next-best method. For distorted waveforms, the optimal MAE drops to 3.9 pm, achieving a 63.38% reduction compared to the next-best method. This study provides both theoretical and technical support for high-precision fiber-optic sensing under complex working conditions. Crucially, the SAK-SVD framework establishes a universal, adaptive denoising paradigm for fiber Bragg grating (FBG) sensing. This paradigm has direct applicability to Raman spectroscopy, industrial ultrasound-based non-destructive testing, and biomedical signal enhancement (e.g., ECG artefact removal), thereby advancing high-precision measurement capabilities across photonics and engineering domains. Full article

Pulsed eddy current (PEC) testing technology has been widely used in the field of non-destructive testing of metal grounding structures due to its wide-band excitation and response characteristics. However, multi-source noise in industrial environments can significantly degrade the performance of PEC sensors, thereby limiting their detection accuracy. This study proposes a multi-modal joint pulsed eddy current signal sensor denoising method that integrates the inductive disturbance mechanism. This method constructs the Improved Whale Optimization -Variational Mode Decomposition-Singular Value Decomposition-Wavelet Threshold Denoising (IWOA-VMD-SVD-WTD) fourth-order processing architecture: IWOA adaptively optimizes the VMD essential variables (K, α) and employs the optimized VMD to decompose the perception coefficient (IMF) of the PEC signal. It utilizes the correlation coefficient criterion to filter and identify the primary noise components within the signal, and the SVD-WTD joint denoising model is established to reconstruct each component to remove the noise signal received by the PEC sensor. To ascertain the efficacy of this approach, we compared the IWOA-VMD-SVD-WTD method with other denoising methods under three different noise levels through experiments. The test results show that compared with other VMD-based denoising techniques, the average signal-to-noise ratio (SNR) of the PEC signal received by the receiving coil for 200 noise signals in different noise environments is 24.31 dB, 29.72 dB and 29.64 dB, respectively. The average SNR of the other two denoising techniques in different noise environments is 15.48 dB, 18.87 dB, 18.46 dB and 19.32 dB, 27.13 dB, 26.78 dB, respectively, which is significantly better than other denoising methods. In addition, in practical applications, this method is better than other technologies in denoising PEC signals and successfully achieves noise reduction and signal feature extraction. This study provides a new technical solution for extracting pure and impurity-free PEC signals in complex electromagnetic environments. Full article

This study analyzes the impact of different types of random noise applied in Denoising Autoencoder (DAE) training on fault diagnosis performance, with the aim of improving noise removal for vibration time series data. While conventional studies typically train DAEs using Gaussian random noise, such noise does not fully reflect the complex noise patterns observed in real-world industrial environments. Therefore, this study proposes a novel approach that uses high-frequency noise components extracted from actual vibration data as training noise for the DAE. Both Gaussian and high-frequency noise were used to train separate DAE models, and statistical features (mean, RMS, standard deviation, kurtosis, skewness) were extracted from the denoised signals. The fault diagnosis rates were calculated using One-Class Support Vector Machines (OC-SVM) for performance comparison. As a result, the model trained with high-frequency noise achieved a 0.0293 higher average F1-score than the Gaussian-based model. Notably, the fault detection accuracy using the kurtosis feature improved significantly from 26.22% to 99.5%. Furthermore, the proposed method outperformed the conventional denoising technique based on the Wavelet Transform, demonstrating superior noise reduction capability. These findings demonstrate that incorporating real high-frequency components from vibration data into the DAE training process is effective in enhancing both noise removal and fault diagnosis performance. Full article

Carbon emission prediction is critical for climate change mitigation across industrial, transportation, and urban sectors. Traditional statistical and machine learning methods struggle to capture complex multi-scale temporal patterns and long-range dependencies in emission data. This paper proposes a hierarchical multi-scale decomposition and deep learning ensemble framework that addresses these limitations. We integrate complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) to decompose carbon emission time series into intrinsic mode functions (IMFs) capturing different frequency bands. Each IMF is processed through a hybrid convolutional neural network (CNN)–Transformer architecture: CNNs extract local features and transformers model long-range dependencies via multi-head attention. An adaptive ensemble mechanism dynamically weights component predictions based on stability and performance metrics. Experiments on four real-world datasets (133,225 observations) demonstrate that our CEEMDAN–CNN–Transformer framework outperforms 12 state-of-the-art methods, achieving a 13.3% reduction in root mean square error (RMSE) to 0.117, 12.7% improvement in mean absolute error (MAE) to 0.088, and 13.0% improvement in continuous ranked probability score (CRPS) to 0.060. The proposed framework not only improves predictive accuracy, but also enhances interpretability by revealing emission patterns across multiple temporal scales, supporting both operational and strategic carbon management decisions. Full article

To address key challenges in clothing trademark quality inspection—namely, insufficient defect samples, unstable performance in complex industrial environments, and low detection efficiency—this paper proposes DLF-YOLO, an enhanced YOLOv11-based model optimized for industrial deployment. To mitigate the problem of limited annotated data, an unsupervised generative network, CycleGAN, is employed to synthesize rare defect patterns and simulate diverse environmental conditions (e.g., rotation, noise, and contrast variations), thereby improving data diversity and model generalization. To reduce the impact of industrial noise, a novel multi-scale dynamic synergy attention (MDSA) attention mechanism is introduced, which utilizes dual attention in both channel and spatial dimensions to focus more accurately on key regions of the trademark, effectively suppressing false detections caused by lighting variations and fabric textures. Furthermore, the high-level selective feature pyramid network (HS-FPN) module is adopted to make the neck structure more lightweight, where the feature selection sub-module enhances the perception of fine edge defects, while the feature fusion sub-module achieves a balance between model lightweighting and detection accuracy through the aggregation of hierarchical multi-scale context information. In the backbone, DWConv replaces standard convolutions before the C3k2 module to reduce computational complexity, and HetConv is integrated into the C3k2 module to simultaneously reduce computational cost and enhance feature extraction capabilities, achieving the goal of maintaining model accuracy. Experimental results on a custom-built dataset demonstrate that DLF-YOLO achieves an mAP@0.5:0.95 of 80.2%, with a 49.6% reduction in parameters and a 25.6% reduction in computational load compared to the original YOLOv11. These results highlight the potential of DLF-YOLO as a scalable and efficient solution for lightweight, industrial-grade defect detection in clothing trademarks. Full article

Developing prefabricated buildings (PBs) and optimizing the construction supply chain represent effective strategies for reducing carbon emissions in the construction industry. Prefabricated building supply chain (PBSC) carbon reduction suffers from synergistic difficulties, limited rationality, and environmental complexity. Therefore, investigating carbon emission reduction in PBSC is essential. In this study, PBSC participants are divided into four categories according to the operation process. Gaussian white noise is introduced to simulate the random perturbation factors, and a four-way stochastic evolutionary game model is constructed and numerically simulated. The study found the following: Stochastic perturbation factors play a prominent role in the evolution speed of the agent; the emission reduction benefit and cost of the participant significantly affect the strategy selection; the operation status of the PBSC is the key to strategy selection, and it is important to pay attention to the synergy of the participants at the first and the last end of the PBSC; the influence of the external environment on strategies is mainly manifested in the loss caused and the assistance provided; and the information on emission reduction is an important factor influencing strategies. Finally, we provide suggestions for promoting carbon emission reduction by participants in the PBSC from the perspective of resisting stochastic perturbation, enhancing participants’ ability, and strengthening PBSC management; external punishment and establishing a cross-industry information sharing platform is more important than the reward. Full article

Investor sentiment has a profound impact on financial market volatility; however, it is difficult to accurately capture the complex nonlinear relationships among sentiment proxies with the existing methods. In this study, we propose a novel investor sentiment indicator, S^GAN, which uses generative adversarial networks (GANs) to extract the nonlinear latent structure from eight sentiment proxies from February 2003 to September 2023 in the Chinese A-share market. Unlike traditional linear dimensionality reduction methods, GANs are able to capture complex market dynamics through adversarial training, effectively reducing noise and improving prediction accuracy. The empirical analyses show that S^GAN significantly outperforms existing methods in both in-sample and out-of-sample prediction capabilities. The GAN-based investment strategy achieves impressive annualized returns and provides a powerful tool for portfolio construction and risk management. Robustness tests across economic cycles, industries, and U.S. markets further validate the stability of S^GAN. These findings highlight the unique advantages of GANs as sentiment-driven financial forecasting tools, providing market participants with new ways to more accurately capture sentiment-shifting trends and develop effective investment strategies. Full article

Balancing ventilation and broadband sound insulation remains a significant challenge in noise control engineering, particularly when simultaneous airflow and broadband noise reduction are required. Conventional porous absorbers and membrane-type metamaterials remain fundamentally constrained by ventilation-blocking configurations or narrow operational bandwidths. This study presents a ventilated composite metamaterial unit (VCMU) co-integrating optimized labyrinth channels and the Helmholtz resonators within a single-plane architecture. This design achieves exceptional ventilation efficiency through a central flow channel while maintaining sub-λ/30 thickness (λ/31 at 860 Hz). Coupled transfer matrix modeling and finite-element simulations reveal that Fano–Helmholtz resonance mechanisms synergistically generate broadband transmission loss (STL) spanning 860–1634 Hz, with six STL peaks in the 860 and 1634 Hz bands (mean 18.4 dB). Experimental validation via impedance tube testing confirmed excellent agreement with theoretical and simulation results. The geometric scalability allows customizable acoustic bandgaps through parametric control. This work provides a promising solution for integrated ventilation and noise reduction, with potential applications in building ventilation systems, industrial pipelines, and other noise-sensitive environments. Full article

Ventilation systems are susceptible to errors, external disruptions, and nonlinear dynamics. Maintaining stable operation and regulating these dynamics require an efficient control system. This study focuses on the speed control of ventilation systems using a super twisted sliding mode observer (STSMO), which provides robust and efficient state estimation for sensorless control. Traditional SM control methods are resistant to parameter fluctuations and external disturbances but are affected by chattering, which degrades performance and can cause mechanical wear. The STSMO leverages the super twisted algorithm, a second-order SM technique, to minimize chattering while ensuring finite-time convergence and high resilience. In sensorless setups, rotor speed and flux cannot be measured directly, making their accurate estimation crucial for effective ventilation drive control. The STSMO enables real-time control by providing current and voltage estimations. It delivers precise rotor flux and speed estimations across varying motor specifications and load conditions using continuous control rules and observer-based techniques. This paper outlines the mathematical formulation of the STSMO, highlighting its noise resistance, chattering reduction, and rapid convergence. Simulation and experimental findings confirm that the proposed observer enhances sensorless ventilation performance, making it ideal for industrial applications requiring reliability, cost-effectiveness, and accuracy. Full article

Air pollution poses a pressing global challenge, particularly in rapidly industrializing nations like China where deteriorating air quality critically endangers public health and sustainable development. To address the heterogeneous patterns of air pollution across diverse geographical and climatic regions, this study proposes a novel CNN-LSTM-KAN hybrid deep learning framework for high-precision Air Quality Index (AQI) time-series prediction. Through systematic analysis of multi-city AQI datasets encompassing five representative Chinese metropolises—strategically selected to cover diverse climate zones (subtropical to temperate), geographical gradients (coastal to inland), and topographical variations (plains to mountains)—we established three principal methodological advancements. First, Shapiro–Wilk normality testing (p < 0.05) revealed non-Gaussian distribution characteristics in the observational data, providing statistical justification for implementing Gaussian filtering-based noise suppression. Second, our multi-regional validation framework extended beyond conventional single-city approaches, demonstrating model generalizability across distinct environmental contexts. Third, we innovatively integrated Kolmogorov–Arnold Networks (KANs) with attention mechanisms to replace traditional fully connected layers, achieving enhanced feature weighting capacity. Comparative experiments demonstrated superior performance with a 23.6–59.6% reduction in Root-Mean-Square Error (RMSE) relative to baseline LSTM models, along with consistent outperformance over CNN-LSTM hybrids. Cross-regional correlation analyses identified PM2.5/PM10 as dominant predictive factors. The developed model exhibited robust generalization capabilities across geographical divisions (R² = 0.92–0.99), establishing a reliable decision-support platform for regionally adaptive air quality early-warning systems. This methodological framework provides valuable insights for addressing spatial heterogeneity in environmental modeling applications. Full article