Search

Long-term solar wind sequence forecasting is essential for understanding the influence of the solar wind on celestial settings, predicting variations in solar wind parameters, and identifying patterns of solar activity. The intrinsic erratic temporal features of solar wind datasets present significant challenges to the development of solar wind factor estimate techniques. In response to these challenges, we present MoCoformer, a novel model based on the Transformer model in deep learning that integrates the Multi-Mode Decomp Block and Mode Independence Attention. The Multi-Mode Decomp Block employs an optimized version of variational mode decomposition technology to flexibly handle irregular features by adaptively decomposing and modeling the impact of sudden events on the temporal dynamics, enhancing its ability to manage non-stationary and irregular features effectively. Meanwhile, the Mode Independence Attention module computes attention independently for each mode, capturing the correlation between sequences and mitigating the negative impact of irregular features on time series prediction. The experimental results on solar wind datasets demonstrate that MoCoformer significantly outperforms current state-of-the-art methods in time series forecasting, showcasing superior predictive performance. This underscores the resilience of MoCoformer in handling the intricate, irregular temporal characteristics of solar wind data, rendering it a valuable instrument for enhancing the understanding and forecasting of solar wind dynamics. [ABSTRACT FROM AUTHOR]

With the widespread application of vortex tube in various fields, it becomes essential to quantitatively explore the separation effect of different fluids (natural fluids, hydrocarbons, etc.) within the vortex tube and to promote its utilization. A new approach has been developed in this study to establish quantitative models for predicting the thermal effects of different fluids in a vortex tube. These models are based on both the macro properties of the working fluid and micro molecular descriptor through a molecular scale. A dataset of 11 numerical simulation results of hydrocarbons and hydrofluorocarbons refrigerants is employed. Three operating conditions, 10 property parameters, and 115 molecular descriptors are screened and identified using random forest feature analysis. Two types of models (micro and macro) have been developed by employing artificial neural network (ANN) modeling techniques. In the result, four key influencing fluid property parameters (the specific heat ratio γ, the multiplication of heat capacity and the molar weight cp·M, the kinematic viscosity ν, and the thermal conductivity λ) and nine molecular descriptors in affecting the thermal effect are identified and respectively chosen as the input in the macro and the micro ANN model establishment. Both types of developed models show a high correlation coefficient (R >.999) and a comparatively low mean square error (MSE). When R600 is employed in the validation, most of the relative error is less than 10%, suggesting both types of models can work effectively in predicting the thermal effect for other fluid. The findings contribute to a deeper understanding of the thermodynamic effect of vortex tubes and provide a valuable tool for selecting and optimizing working fluids in various applications. [ABSTRACT FROM AUTHOR]

Background: Currently, it is still confused whether preoperative aminotransferase-to-platelet ratio (APRI) and gamma-glutamyl transferase-to-platelet ratio (GPR) can predict microvascular invasion (MVI) in solitary hepatocellular carcinoma (HCC). We aimed to develop and validate a machine-learning integration model for predicting MVI using APRI, GPR and gadoxetic acid disodium (Gd-EOB-DTPA) enhanced MRI. Methods: A total of 314 patients from XinQiao Hospital of Army Medical University were divided chronologically into training set (n = 220) and internal validation set (n = 94), and recurrence-free survival was determined to follow up after surgery. Seventy-three patients from Chongqing University Three Gorges Hospital and Luzhou People's Hospital served as external validation set. Overall, 387 patients with solitary HCC were analyzed as whole dataset set. Least absolute shrinkage and selection operator, tenfold cross-validation and multivariate logistic regression were used to gradually filter features. Six machine-learning models and an ensemble of the all models (ENS) were built. The area under the receiver operating characteristic curve (AUC) and decision curve analysis were used to evaluate model's performance. Results: APRI, GPR, HBPratio3 ([liver SI‒tumor SI]/liver SI), PLT, peritumoral enhancement, non-smooth margin and peritumoral hypointensity were independent risk factors for MVI. Six machine-learning models showed good performance for predicting MVI in training set (AUCs range, 0.793– 0.875), internal validation set (0.715– 0.832), external validation set (0.636– 0.746) and whole dataset set (0.756– 0.850). The ENS achieved the highest AUCs (0.879 vs 0.858 vs 0.839 vs 0.851) in four cohorts with excellent calibration and more net benefit. Subgroup analysis indicated that ENS obtained excellent AUCs (0.900 vs 0.809 vs 0.865 vs 0.908) in HCC > 5cm, ≤ 5cm, ≤ 3cm and ≤ 2cm cohorts. Kaplan‒Meier survival curves indicated that ENS achieved excellent stratification for MVI status. Conclusion: The APRI and GPR may be new potential biomarkers for predicting MVI of HCC. The ENS achieved optimal performance for predicting MVI in different sizes HCC and may aid in the individualized selection of surgical procedures. [ABSTRACT FROM AUTHOR]

The article focuses on a method for monitoring soil moisture content (SMC) and temperature using hyperspectral technology. Topics discussed include the application of successive projection algorithm (SPA) for extracting characteristic wavelengths, and optimizing hyperparametric weight and bias values through genetic algorithm (GA) for support vector machine regression (SVR). It also mentions a SPA-GA-SVR hybrid model that outperforms other models in predicting SMC and temperature.

Purpose: Osteoarthritis (OA) is a joint disease with a long and slow course, which is one of the major causes of disability in middle and old-aged people. This study was dedicated to excavating the cellular senescence-associated biomarkers of OA.Methods: The Gene Expression Omnibus (GEO) database was searched and five datasets pertaining to OA were obtained. After removing the batch effect, the GSE55235, GSE55457, GSE82107, and GSE12021 datasets were integrated together for screening of the candidate genes by differential analysis and weighted gene co-expression network analysis (WGCNA). Next, those genes were further filtered by machine learning algorithms to obtain cellular senescence-associated biomarkers of OA. Subsequently, enrichment analyses based on those biomarkers were conducted, and we profiled the infiltration levels of 22 types immune cells with the ERSORT algorithm. A lncRNA-miRNA-mRNA regulatory and drug-gene network were constructed. Finally, we validated the senescence-associated biomarkers at both in vivo and in vitro levels.Results: Five genes (BCL6, MCL1, SLC16A7, PIM1, and EPHA3) were authenticated as cellular senescence-associated biomarkers in OA. ROC curves demonstrated the reliable capacity of the five genes as a whole to discriminate OA samples from normal samples. The nomogram diagnostic model based on 5 genes proved to be a reliable predictor of OA. Single-gene GSEA results pointed to the involvement of the five biomarkers in immune-related pathways and oxidative phosphorylation in the development of OA. Immune infiltration analysis manifested that the five genes were significantly correlated with differential immune cells. Subsequently, a lncRNA-miRNA-mRNA network and gene-drug network containing were generated based on five cellular senescence-associated biomarkers in OA.Conclusion: A foundation for understanding the pathophysiology of OA and new insights into OA diagnosis and treatment were provided by the identification of five genes, namely BCL6, MCL1, SLC16A7, PIM1, and EPHA3, as biomarkers associated with cellular senescence in OA. [ABSTRACT FROM AUTHOR]

Liver cirrhosis remains a significant global public health concern, with liver transplantation standing as the foremost effective treatment currently available. Therefore, investigating the pathogenesis of liver cirrhosis and developing novel therapies is imperative. Mitochondrial dysfunction stands out as a pivotal factor in its development. This study aimed to elucidate the relationship between mitochondria dysfunction and liver cirrhosis using bioinformatic methods to unveil its pathogenesis. Initially, we identified 460 co-expressed differential genes (co-DEGs) from the GSE14323 and GSE25097 datasets, alongside their combined datasets. Functional analysis revealed that these co-DEGs were associated with inflammatory cytokines and cirrhosis-related signaling pathways. Utilizing weighted gene co-expression network analysis (WCGNA), we screened module genes, intersecting them with co-DEGs and oxidative stress-related mitochondrial genes. Two algorithms (least absolute shrinkage and selection operator (LASSO) regression and SVE-RFE) were then employed to further analyze the intersecting genes. Finally, COX7A1 and IFI27 emerged as identifying genes for liver cirrhosis, validated through a receiver operating characteristic (ROC) curve analysis and related experiments. Additionally, immune infiltration highlighted a strong correlation between macrophages and cirrhosis, with the identifying genes (COX7A1 and IFI27) being significantly associated with macrophages. In conclusion, our findings underscore the critical role of oxidative stress-related mitochondrial genes (COX7A1 and IFI27) in liver cirrhosis development, highlighting their association with macrophage infiltration. This study provides novel insights into understanding the pathogenesis of liver cirrhosis. [ABSTRACT FROM AUTHOR]

• N-MBL was compared with Machine Learning and local models. • N-MBL explained the non-linear relationship well between the property and spectra. • The performance of N-MBL was better than Machine Learning models. • The number of selected nearest neighbors was related to the accuracy of local model. Visible near-infrared (vis-NIR) spectroscopy has gained widespread recognition as an efficient and reliable approach for the rapid monitoring of soil properties. This technique relies on robust machine learning models that convert soil spectra information to soil properties. In particular, memory-based learning (MBL) has emerged as a powerful local modeling technique for soil spectral analysis. However, conventional MBL algorithms use linear models, disregarding the non-linear relationship between soil properties and vis-NIR spectra. Therefore, we hypothesize that non-linear memory-based learning (N-MBL) models can enhance prediction. This study develops and evaluates the N-MBL algorithm using the Lateritic Red soil spectral library (LRSSL) from Guangdong province in China. This library consists of 742 samples of vis-NIR spectra and corresponding soil properties, including pH, soil organic matter (SOM), total nitrogen (TN), total phosphorus (TP), and total potassium (TK). As a comparison, several commonly used supervised learning methods, such as Partial least squares regression (PLSR), Cubist, Random Forest (RF), Super vector machine (SVM), Convolution neural network (CNN), and local models (MBL), were compared to the proposed N-MBL. The results showed that local models generally outperformed supervised learning methods, particularly when applied to a large soil spectral library with a substantial number of samples (over 500). When comparing the two local models, MBL had more fluctuation of model performance compared to N-MBL as the number of selected nearest neighbors (k) varied between 30 and 250. As k increased, N-MBL showed higher R2 values for SOM and TN prediction than MBL but lower performance for pH and TK prediction. In addition, N-MBL outperformed MBL in predicting TP. In conclusion, N-MBL is a new local algorithm for predicting soil properties from vis-NIR spectra. It has a high potential to improve the accuracy of the prediction of multiple soil properties. [ABSTRACT FROM AUTHOR]

A bionic tactile device is designed for object surface texture recognition, taking inspiration from the microstructure of human fingerprints. The sense of touch in humans is achieved through the frictional vibration and generation of electrical potential signals by subcutaneous receptors. Constructing a tactile sensing device involves generating distinct signals upon interacting with different materials. The piezoelectric film PVDF is particularly suitable as a sensitive material for sensors due to its excellent flexibility, strong mechanical strength, excellent dynamic response and cost-effectiveness. This paper presents the design of a PVDF-based fingerprint-inspired tactile sensor capable of differentiating various textures. By combining the collected signals with machine learning algorithms, diverse textures can be effectively identified. To demonstrate the sensor's superior performance, two experiments were conducted—one focused on recognizing different material textures, and the other on recognizing Braille characters. The accuracy achieved in these experiments was 97.4% and 96.5%, respectively, highlighting the technology's significant potential in intelligent robotics and human-computer interaction. [Display omitted] • A bionic tactile sensor with three different layers were proposed. • Two types of texture datasets were representatively constructed for analytical study. • Some eigenvalue parameters combined with machine learning algorithms was used to achieve high recognition accuracy. [ABSTRACT FROM AUTHOR]

The inspection of broken wires in prestressed concrete cylinder pipes is crucial for ensuring the safety and reliability of the pipeline. Traditional point detection techniques always require labor-intensive periodic inspections and cannot deployed along the entire pipeline, significantly limiting the development of the industry. Hence, there is an urgent need for more advanced and intelligent sensors that can achieve 100% coverage and provide sufficient accuracy assurance. In this work, we develop a distributed acoustic sensing -based automated monitoring system to accurately classify the rupture of prestressed wires. First, a computer vision approach is employed to primarily screen out potential vibrational signals from DAS array images. Then, a pre-trained support vector machine model is used to classify the vibrations as either wire breakages or non-wire breakages. This model's performance surpassed other classification strategies, achieving 99.62% accuracy, 99.41% precision, 98.82% recall, and 99.12% F1-score in a side-to-side comparison. Our innovative workflow provides a comprehensive solution for detecting broken wires and offers guidance for the application of artificial intelligence-based DAS to complex vibration systems with limited training data. [ABSTRACT FROM AUTHOR]

Collaborative fairness in federated learning rewards high-contribution clients with high-performance models when multiple clients train a machine learning model cooperatively. Existing approaches ignore the information on data distribution when evaluating the clients' data quality, resulting in a mismatch between the reward allocation and the real data quality of clients under different data heterogeneity settings. To address this problem, we propose a novel Federated learning framework with Adaptive data Value Evaluation mechanism (FedAVE) to ensure collaborative fairness without affecting the predictive performance of models. First, an adaptive reputation calculation module is designed to generate reputations that match the clients' contributions based on the information of their data distribution, respectively. Second, a dynamic gradient reward distribution module is devised to allocate a certain number of aggregated model parameter updates/gradients as the rewards corresponding to the reputations and the data distribution information. Extensive experiments on three public benchmarks show that the proposed FedAVE outperforms all baseline methods in terms of fairness, and achieves comparable performance to the state-of-the-art methods in terms of accuracy. Code available at https://github.com/wangzihuixmu/FedAVE. [ABSTRACT FROM AUTHOR]