Search

Background: Early recognition and timely intervention for AKI in critically ill patients were crucial to reduce morbidity and mortality. This study aimed to use biomarkers to construct a optimal machine learning model for early prediction of AKI in critically ill patients within seven days., Methods: The prospective cohort study enrolled 929 patients altogether who were admitted in ICU including 680 patients in training set (Jiefang Campus) and 249 patients in external testing set (Binjiang Campus). After performing strict inclusion and exclusion criteria, 421 patients were selected in training set for constructing predictive model and 167 patients were selected in external testing for evaluating the predictive performance of resulting model. Urine and blood samples were collected for kidney injury associated biomarkers detection. Baseline clinical information and laboratory data of the study participants were collected. We determined the average prediction efficiency of six machine learning models through 10-fold cross validation., Results: In total, 78 variables were collected when admission in ICU and 43 variables were statistically significant between AKI and non-AKI cohort. Then, 35 variables were selected as independent features for AKI by univariate logistic regression. Spearman correlation analysis was used to remove two highly correlated variables. Three ranking methods were used to explore the influence of 33 variables for further determining the best combination of variables. The gini importance ranking method was found to be applicable for variables filtering. The predictive performance of AKIML pred which constructed by the XGBoost algorithm was the best among six machine learning models. When the AKIML pred included the nine features (NGAL, IGFBP7, sCysC, CAF22, KIM-1, NT-proBNP, IL-6, IL-18 and L-FABP) with the highest influence ranking, its model had the best prediction performance, with an AUC of 0.881 and an accuracy of 0.815 in training set, similarly, with an AUC of 0.889 and an accuracy of 0.846 in validation set. Moreover, the performace was slightly outperformed in testing set with an AUC of 0.902 and an accuracy of 0.846. The SHAP algorithm was used to interpret the prediction results of AKIML pred . The web-calculator of AKIML pred was shown for predicting AKI with more convenient(https://www.xsmartanalysis.com/model/list/predict/model/html?mid=8065&symbol=11gk693982SU6AE1ms21). AKIML pred was better than the optimal model built with only routine tests for predicting AKI in critically ill patients within 7 days., Conclusion: The model AKIML pred constructed by the XGBoost algorithm with selecting the nine most influential biomarkers in the gini importance ranking method had the best performance in predicting AKI in critically ill patients within 7 days. This data-driven predictive model will help clinicians to make quick and accurate diagnosis., (Copyright © 2024 Elsevier B.V. All rights reserved.)

Exabytes of data are generated daily by humans, leading to the growing needs for new efforts in dealing with the grand challenges for multi-label learning brought by big data. For example, extreme multi-label classification is an active and rapidly growing research area that deals with classification tasks with extremely large number of classes or labels; utilizing massive data with limited supervision to build a multi-label classification model becomes valuable for practical applications, etc. Besides these, there are tremendous efforts on how to harvest the strong learning capability of deep learning to better capture the label dependencies in multi-label learning, which is the key for deep learning to address real-world classification tasks. However, it is noted that there have been a lack of systemic studies that focus explicitly on analyzing the emerging trends and new challenges of multi-label learning in the era of big data. It is imperative to call for a comprehensive survey to fulfil this mission and delineate future research directions and new applications.

The evergrowing diversity of encrypted and anonymous network traffic makes network management more formidable to manage the network traffic. An intelligent system is essential to analyse and identify network traffic accurately. Network management needs such techniques to improve the Quality of Service and ensure the flow of secure network traffic. However, due to the usage of non‐standard ports and encryption of data payloads, the classical port‐based and payload‐based classification techniques fail to classify the secured network traffic. To solve the above‐mentioned problems, this paper proposed an effective deep learning‐based framework employed with flow‐time‐based features to predict heterogeneous secure network traffic best. The state‐of‐the‐art machine learning strategies (C4.5, random forest, and K‐nearest neighbour) are investigated for comparison. The proposed 1D‐CNN model achieved higher accuracy in classifying the heterogeneous secure network traffic. In the next step, the proposed deep learning model characterises the major categories (virtual private network traffic, the onion router network traffic, and plain encrypted network traffic) into several application types. The experimental results show the effectiveness and feasibility of the proposed deep learning framework, which yields improved predictive power compared to the state‐of‐the‐art machine learning techniques employed for secure network traffic analysis. [ABSTRACT FROM AUTHOR]

Every year, a large amount of solid waste such as fly ash and slag is generated worldwide. If these solid wastes are used in concrete mixes to make concrete, it can effectively save resources and protect the environment. The compressive strength of concrete is an essential indicator for testing its quality, and its prediction is affected by many factors. It is difficult to predict its strength accurately. Therefore, based on the current popular machine learning supervised learning algorithms: Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Support Vector Machine (SVR), three models established a nonlinear mapping between multi-factor features and target feature concrete compressive strength. Using the three completed training models, we validated the test set with 206 example sets, and the Root Mean Square Error (RMSE), fitting coefficient (R2), and Mean Absolute Error (MAE) were used as evaluation metrics. The validation results showed that the values of RMSE, R2, and MAE for the RF model were 0.1, 0.9, and 0.21, respectively; the values of XGBoost model were 0.05, 0.95, and 0.15, respectively. The values of SVR were 0.15, 0.86, and 0.3, respectively. As a result, Extreme Gradient Boosting (XGBoost) has better generalization ability and prediction accuracy than the other two algorithms. [ABSTRACT FROM AUTHOR]

Background: Predicting amnestic mild cognitive impairment (aMCI) in conversion and Alzheimer's disease (AD) remains a daunting task. Standard diagnostic procedures for AD population are reliant on neuroimaging features (positron emission tomography, PET), cerebrospinal fluid (CSF) biomarkers (Aβ1-42, T-tau, P-tau), which are expensive or require invasive sampling. The blood-based biomarkers offer the opportunity to provide an alternative approach for easy diagnosis of AD, which would be a less invasive and cost-effective screening tool than currently approved CSF or amyloid β positron emission tomography (PET) biomarkers. Methods: We developed and validated a sensitive and selective immunoassay for total Tau in plasma. Robust signatures were obtained based on several clinical features selected by multiple machine learning algorithms between the three participant groups. Subsequently, a well-fitted nomogram was constructed and validated, integrating clinical factors and total Tau concentration. The predictive performance was evaluated according to the receiver operating characteristic (ROC) curves and area under the curve (AUC) statistics. Decision curve analysis and calibration curves are used to evaluate the net benefit of nomograms in clinical decision-making. Results: Under optimum conditions, chemiluminescence analysis (CLIA) displays a desirable dynamic range within Tau concentration from 7.80 to 250 pg/mL with readily achieved higher performances (LOD: 5.16 pg/mL). In the discovery cohort, the discrimination between the three well-defined participant groups according to Tau concentration was in consistent agreement with clinical diagnosis (AD vs. non-MCI: AUC = 0.799; aMCI vs. non-MCI: AUC = 0.691; AD vs. aMCI: AUC = 0.670). Multiple machine learning algorithms identified Age, Gender, EMPG, Tau, ALB, HCY, VB12, and/or Glu as robust signatures. A nomogram integrated total Tau concentration and clinical factors provided better predictive performance (AD vs. non-MCI: AUC = 0.960, AD vs. aMCI: AUC = 0.813 in discovery cohort; AD vs. non-MCI: AUC = 0.938, AD vs. aMCI: AUC = 0.754 in validation cohort). Conclusion: The developed assay and a satisfactory nomogram model hold promising clinical potential for early diagnosis of aMCI and AD participants. Flow diagram of the study design. [ABSTRACT FROM AUTHOR]

Image scene classification aiming to assign specific semantic labels for each image is vitally important for the applications of remote sensing (RS) data. In real world, since the observation environment is open and dynamic, RS images are collected sequentially and the numbers of images and classes grow rapidly over time. Most existing scene classification methods are offline learning algorithms, which are inefficient and unscalable for this scenario. In this article, an incremental learning with open-set recognition (ILOSR) framework is proposed for RS image scene classification in the open and dynamic environment, which can identify the unknown classes from a stream of data and learn these new classes incrementally. Specifically, a controllable convex hull-based exemplar selection strategy is designed to address the catastrophic forgetting issue in incremental learning, which can reduce training time and memory footprint effectively. In addition, a new loss function based on prototype learning and uncertainty measurement is proposed for OSR to enhance the interclass discrimination and intraclass compactness of the learned deep features. Experimental results on real RS datasets demonstrate that the proposed method can not only outperform the state-of-the-art approaches on offline classification, incremental learning, and OSR problem separately but also achieve better and more stable performance in the experiments for ILOSR. [ABSTRACT FROM AUTHOR]

The comprehensive properties of high-entropy alloys (HEAs) are highly-dependent on their phases. Although a large number of machine learning (ML) algorithms has been successfully applied to the phase prediction of HEAs, the accuracies among different ML algorithms based on the same dataset vary significantly. Therefore, selection of an efficient ML algorithm would significantly reduce the number and cost of the experiments. In this work, phase prediction of HEAs (PPH) is proposed by integrating criterion and machine learning recommendation method (MLRM). First, a meta-knowledge table based on characteristics of HEAs and performance of candidate algorithms is established, and meta-learning based on the meta-knowledge table is adopted to recommend an algorithm with desirable accuracy. Secondly, an MLRM based on improved meta-learning is engineered to recommend a more desirable algorithm for phase prediction. Finally, considering poor interpretability and generalization of single ML algorithms, a PPH combining the advantages of MLRM and criterion is proposed to improve the accuracy of phase prediction. The PPH is validated by 902 samples from 12 datasets, including 405 quinary HEAs, 359 senary HEAs, and 138 septenary HEAs. The experimental results shows that the PPH achieves performance than the traditional meta-learning method. The average prediction accuracy of PPH in all, quinary, senary, and septenary HEAs is 91.6%, 94.3%, 93.1%, and 95.8%, respectively. [ABSTRACT FROM AUTHOR]

Background: The non-invasive preoperative diagnosis of microvascular invasion (MVI) in hepatocellular carcinoma (HCC) is vital for precise surgical decision-making and patient prognosis. Herein, we aimed to develop an MVI prediction model with valid performance and clinical interpretability. Methods: A total of 2160 patients with HCC without macroscopic invasion who underwent hepatectomy for the first time in West China Hospital from January 2015 to June 2019 were retrospectively included, and randomly divided into training and a validation cohort at a ratio of 8:2. Preoperative demographic features, imaging characteristics, and laboratory indexes of the patients were collected. Five machine learning algorithms were used: logistic regression, random forest, support vector machine, extreme gradient boosting (XGBoost), and multilayer perception. Performance was evaluated using the area under the receiver operating characteristic curve (AUC). We also determined the Shapley Additive exPlanation value to explain the influence of each feature on the MVI prediction model. Results: The top six important preoperative factors associated with MVI were the maximum image diameter, protein induced by vitamin K absence or antagonist-II, α-fetoprotein level, satellite nodules, alanine aminotransferase (AST)/aspartate aminotransferase (ALT) ratio, and AST level, according to the XGBoost model. The XGBoost model for preoperative prediction of MVI exhibited a better AUC (0.8, 95% confidence interval: 0.74–0.83) than the other prediction models. Furthermore, to facilitate use of the model in clinical settings, we developed a user-friendly online calculator for MVI risk prediction based on the XGBoost model. Conclusions: The XGBoost model achieved outstanding performance for non-invasive preoperative prediction of MVI based on big data. Moreover, the MVI risk calculator would assist clinicians in conveniently determining the optimal therapeutic remedy and ameliorating the prognosis of patients with HCC. [ABSTRACT FROM AUTHOR]

While many organizations' hiring practices now incorporate artificial intelligence (AI) and machine learning (ML), research suggests that job applicants may react negatively toward AI/ML-based selection practices. In the current research, we thus examined how organizations might mitigate adverse reactions toward AI/ML-based selection processes. In two between-subjects experiments, we recruited online samples of participants (undergraduate students and Prolific panelists, respectively) and presented them with vignettes representing various selection systems and measured participants' reactions to them. In Study 1, we manipulated (a) whether the system was managed by a human decision-maker, by AI/ML, or a combination of both (an "augmented" approach), and (b) the selection stage (screening, final stage). Results indicated that participants generally reacted more favorably toward augmented and human-based approaches, relative to AI/ML-based approaches, and further depended on participants' pre-existing familiarity levels with AI. In Study 2, we sought to replicate our findings within a specific process (selecting hotel managers) and application method (handling interview recordings). We found again that reactions toward the augmented approach generally depended on participants' familiarity levels with AI. Our findings have implications for how (and for whom) organizations should implement AI/ML-based practices. • People react more favorably toward AI that augments (rather than replaces) humans. • Reactions to AI are relevant to both early and late stages of the hiring process. • Augmented approaches mitigate adverse reactions to AI-based hiring for people who have low familiarity with AI. [ABSTRACT FROM AUTHOR]

• DenseNet has better prediction performance than AlexNet, GoogLeNet, and ResNets. • DenseNet-39 gained 99.3% accuracy with low computational cost and high efficiency. • t-SNE and confusion matrix was used to visualize the performance of DenseNet-39. • CAM was used to explore the mechanism of false identification by DenseNet-39. Detecting and classifying the melt pool states in laser-based direct energy deposition (L-DED) is crucial for reducing defects and enhancing the mechanical properties of L-DED metal parts. Although physics-based modeling methods and traditional machine learning algorithms such as convolutional neural network have been introduced to monitor the melt pool states, improving the low accuracy of these methods remains to be challenging. To address this issue, we developed a DenseNet-39 model to classify the melt pool states. 80 single-track samples were fabricated using a linear scan strategy by L-DED and using a coaxial high-speed camera to capture the melt pool images in-process. Experimental results have demonstrated the superior performance of DenseNet-39 in classifying the melt pool states with 99.3% accuracy, achieved a lower computation burden, and less processing time. We used CAM to explain the mechanism of classification by DenseNet-39. DenseNet-39 provides the potential applications of online process monitoring in L-DED. [ABSTRACT FROM AUTHOR]

RNA polymerase II (RNA Pol II) has been recognized as a passively regulated multi-subunit holoenzyme. However, the extent to which RNA Pol II subunits might be important beyond the RNA Pol II complex remains unclear. Here, fractions containing disassociated RPB3 (dRPB3) were identified by size exclusion chromatography in various cells. Through a unique strategy, i.e., "specific degradation of disassociated subunits (SDDS)," we demonstrated that dRPB3 functions as a regulatory component of RNA Pol II to enable the preferential control of 3′ end processing of ribosomal protein genes directly through its N-terminal domain. Machine learning analysis of large-scale genomic features revealed that the little elongation complex (LEC) helps to specialize the functions of dRPB3. Mechanistically, dRPB3 facilitates CBC-PCF11 axis activity to increase the efficiency of 3′ end processing. Furthermore, RPB3 is dynamically regulated during development and diseases. These findings suggest that RNA Pol II gains specific regulatory functions by trapping disassociated subunits in mammalian cells. [Display omitted] • Disassociated RPB3 (dRPB3) exists in various cells • dRPB3 preferentially controls 3′ end processing of ribosomal subunit genes • The RPB3 N terminus directly interacts with CBC and is required for 3′ end processing • LEC is associated with the specificity of the dRPB3-mediated CBC-PCF11 axis Li et al. show that disassociated RPB3 imparts the preferential regulation of 3′ end processing for ribosomal subunit genes. This indicates that RNA Pol II subunits cause RNA Pol II to transform from a mono-functional enzyme to a multi-functional enzyme, increasing enzyme efficiency and conferring specificity regulations. [ABSTRACT FROM AUTHOR]