Search

Background: Osteoporosis is becoming more common worldwide, imposing a substantial burden on individuals and society. The onset of osteoporosis is subtle, early detection is challenging, and population-wide screening is infeasible. Thus, there is a need to develop a method to identify those at high risk for osteoporosis., Objective: This study aimed to develop a machine learning algorithm to effectively identify people with low bone density, using readily available demographic and blood biochemical data., Methods: Using NHANES 2017-2020 data, participants over 50 years old with complete femoral neck BMD data were selected. This cohort was randomly divided into training (70%) and test (30%) sets. Lasso regression selected variables for inclusion in six machine learning models built on the training data: logistic regression (LR), support vector machine (SVM), gradient boosting machine (GBM), naive Bayes (NB), artificial neural network (ANN) and random forest (RF). NHANES data from the 2013-2014 cycle was used as an external validation set input into the models to verify their generalizability. Model discrimination was assessed via AUC, accuracy, sensitivity, specificity, precision and F1 score. Calibration curves evaluated goodness-of-fit. Decision curves determined clinical utility. The SHAP framework analyzed variable importance., Results: A total of 3,545 participants were included in the internal validation set of this study, of whom 1870 had normal bone density and 1,675 had low bone density Lasso regression selected 19 variables. In the test set, AUC was 0.785 (LR), 0.780 (SVM), 0.775 (GBM), 0.729 (NB), 0.771 (ANN), and 0.768 (RF). The LR model has the best discrimination and a better calibration curve fit, the best clinical net benefit for the decision curve, and it also reflects good predictive power in the external validation dataset The top variables in the LR model were: age, BMI, gender, creatine phosphokinase, total cholesterol and alkaline phosphatase., Conclusion: The machine learning model demonstrated effective classification of low BMD using blood biomarkers. This could aid clinical decision making for osteoporosis prevention and management., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2024 Xu, Chen, Yao, Wu, Cui, Wang, Diao, Jin, Hong and Li.)

Cell fate decisions and transitions are common in almost all developmental processes. Therefore, it is important to identify the decision-making mechanisms and important individual molecules behind the fate decision processes. In this paper, we propose an interpretable strategy based on systematic perturbation, unsupervised hierarchical cluster analysis (HCA), machine learning (ML), and Shapley additive explanation (SHAP) analysis for inferring the contribution and importance of individual molecules in cell fate decision and transition processes. In order to verify feasibility of the approach, we apply it to the core epithelial to mesenchymal transition (EMT)-metastasis network. The key factors identified in EMT-metastasis are consistent with relevant experimental observations. The approach presented here can be applied to other biological networks to identify important factors related to cell fate decisions and transitions., (© 2023. The Author(s), under exclusive licence to Springer Nature B.V.)

Chloride ion corrosion has been considered to be one of the main reasons for durability deterioration of reinforced concrete structures in marine or chlorine-containing deicing salt environments. This paper studies the relationship between the amount of fly ash and the durability of concrete, especially the resistance to chloride ion erosion. The heat trend map of total chloride ion factor correlation displayed that the ranking of factor correlations was as follows: sampling depth > cement dosage > fly ash dosage. In order to verify the effect of fly ash dosage on chloride ion resistance, three different machine learning algorithms (RF, GBR, DT) are employed to predict the total chloride content of fly ash proportioned concrete with varying admixture ratios, which are evaluated based on R2, MSE, RMSE, and MAE. The results predicted by the RF model show that the threshold of fly ash admixture in chlorinated salt environments is 30–40%. Replacing part of cement with fly ash in the mixture of concrete below this threshold of fly ash, it could change the phase structure and pore structure, which could improve the permeability of fly ash concrete and reduce the content of free chloride ions in the system. Machine learning modeling using sample data can accurately predict concrete properties, which effectively reduce engineering tests. The development of machine learning models is essential for the decarbonization and intelligence of engineering. [ABSTRACT FROM AUTHOR]

Intelligence modeling interpretable for incipient fault diagnosis of batch process represents a serious challenge. For the multivariate, nonlinear, and high-dimensionality characteristics of process data, existing fault diagnosis solutions are easily concealed by noise while neglect the low amplitude and noise interference of the incipient faults. In this paper, an intelligent incipient fault diagnosis model based on deep time series feature extraction network is proposed, which integrates denoising autoencoder (DAE), layer normalization and dropout layer added LSTM (LD-LSTM), and stacked autoencoder (SAE) to obtain the efficient features. DAE is applied to restore the data and eliminate noise interference. LD-LSTM is utilized to extract time series features, in which the layer normalization and the dropout layer are added between the layers of the LSTM to solve the problems of slow model convergence and over-fitting. SAE performs the feature extraction to obtain deep time series features and learns more efficient expressions. The softmax function is applied to achieve the incipient fault diagnosis based on the deep time series features. A case study on a fed-batch penicillin fermentation process suggests that the proposed method can obtain an accuracy rate of 98.97%, which has increased an average of 20% than the traditional shallow machine learning. The accuracy of fault recognition and comparative experiments demonstrate the effectiveness and feasibility of this model, and provide a solution for the application of intelligent fault diagnosis in batch industries. [ABSTRACT FROM AUTHOR]

Many existing methods formulate the action prediction task as recognizing early parts of actions in trimmed videos. In this paper, we focus on predicting actions from ongoing untrimmed videos where actions might not happen at the very beginning of videos. It is extremely challenging to predict actions in such untrimmed videos due to ambiguous or even no information of actions in the early parts of videos. To address this problem, we propose a prediction confidence that assesses the decision quality of a prediction model. Guided by the confidence, the model continuously refines the prediction results by itself with the increasing observed video frames. Specifically, we build a Self Prediction Refining Network (SPR-Net) which incrementally learns the confidence for action prediction. SPR-Net consists of three modules: a temporal hybrid network, an incremental confidence learner, and a self-refining Gumbel softmax sampler. The temporal hybrid network generates the action category distributions by integrating static scene and dynamic motion information. The incremental confidence learner calculates the confidence in an incremental manner, judging the extent to which the temporal hybrid network should believe its prediction result. The self-refining Gumbel softmax sampler models the mutual relationship between the prediction confidence and the category distribution, which enables them to be jointly learned in an end-to-end fashion. We also present a sparse self-attention mechanism to encode local spatio-temporal features into the frame-level motion representation to further improve the prediction performance. Extensive experiments on five datasets (i.e., UT-Interaction, BIT-Interaction, UCF101, THUMOS14, and ActivityNet) validate the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

With the accumulation of data on personal behavior and the development of machine learning models and algorithms, it is becoming possible to use social media data for personal credit scoring. In this article, we use the systematic sampling method to obtain Douban's social media data. Because there are many abnormal users in these data, they are "real but false data" for personal credit evaluation. In order to better carry out personal credit scoring, we propose three criteria, power exponents of time interval distribution of individual user $\gamma _i$γi, user activity $A_i$Ai, and the ratio of out-degree and in-degree $R_i$Ri of user $i$i, which are used to systematically clean the data. And then, we used the logistic regression method to score the individual credits of users before and after data cleaning, and found that the rank order of personal credit scoring has changed significantly. This change is largely attributed to the changes of network structure after data cleaning. We believe that our work is very important to use the social media data to establish a credible personal credit evaluation system to reduce the credit risk of the current Internet financial industry. [ABSTRACT FROM AUTHOR]

Signal loss in blood oxygen level-dependent (BOLD) functional neuroimaging is common and can lead to misinterpretation of findings. Here, we reconstructed compromised fMRI signal using deep machine learning. We trained a model to learn principles governing BOLD activity in one dataset and reconstruct artificially compromised regions in an independent dataset, frame by frame. Intriguingly, BOLD time series extracted from reconstructed frames are correlated with the original time series, even though the frames do not independently carry any temporal information. Moreover, reconstructed functional connectivity maps exhibit good correspondence with the original connectivity maps, indicating that the model recovers functional relationships among brain regions. We replicated this result in two healthy datasets and in patients whose scans suffered signal loss due to intracortical electrodes. Critically, the reconstructions capture individual-specific information. Deep machine learning thus presents a unique opportunity to reconstruct compromised BOLD signal while capturing features of an individual's own functional brain organization. Signal loss in blood oxygen level‐dependent (BOLD) fMRI can lead to misinterpretation of findings. The authors trained a deep learning model to reconstruct compromised BOLD signal in datasets from healthy participants and in patients whose scans suffered signal loss due to intracortical electrodes. [ABSTRACT FROM AUTHOR]

This article offers a detailed investigation into the technical, economic along with environmental performance of four configurations of hybrid renewable energy systems (HRESs), aiming at supplying renewable electricity to a remote location, Henry Island in India. The study explores combinations involving photovoltaic (PV) panels, wind turbines, biogas generators, batteries, and converters, while evaluating their economic, technical, and environmental performance. The economic analysis yield that among all the systems examined, the PV, wind turbine, biogas generator, battery, and converter integrated configuration stands out with highly favourable results, showcasing the minimal value of levelized cost of electricity (LCOE) at $0.4224 per kWh and the lowest net present cost (NPC) at $6.41 million. However, technical analysis yield that the configuration comprising wind turbines, PV panels, converters, and battery yields a maximum excess electricity output of 2,838,968 kWh/yr. Additionally, machine learning techniques are employed to analyse economic and environmental performance data. The study shows Bilayered Neural Network model achieves exceptional accuracy in predicting LCOE, while the Medium Neural Network model proves to be the most accurate in predicting environmental performance. These findings provide valuable perception into the design and optimisation of HRES systems for off-grid applications in remote regions, taking into account their technical, economic, and environmental aspects. • Four different configurations of hybrid renewable energy systems are considered. • The configuration that integrated PV, wind turbine, biogas generator, battery, and converter is best. • Machine learning techniques are used to assess economic and environmental performances. • The bilayered neural network, with ReLU activation function, outperforms other models in predicting LCOE with R2 = 1 • The medium neural network, using ReLU activation, outperforms other models in predicting CO 2 emissions with R2 = 1. [ABSTRACT FROM AUTHOR]