Search

Air quality models are increasingly important in air pollution forecasting and control. Sectoral emissions significantly impact the accuracy of air quality models and source apportionment. This paper studied the 3DVar (three-dimensional variational) emission inversion method, which is based on machine learning, and then expanded it to sectoral emission inversion combined with source apportionment. Two machine learning conversion matrices were established to implement this method: a matrix that converts the total pollutant concentration to sectoral source apportionment results and a matrix that converts the sectoral source apportionment results to corresponding emissions. Combined with the O 3 (ozone) concentration contributed by VOCs (volatile organic compounds) and NO x (nitrogen oxides) precursors in source apportionment, the inversion ability for O 3 -NO x -VOCs nonlinear processes was improved. Taking the BTH (Beijing‒Tianjin-Hebei) region from January 15 to 30, 2019, as an example, the results revealed that the regional errors of PM 2.5 and O 3 in the inversion experiment were reduced by 47% and 45%, respectively, and the temporal errors were reduced by 44% and 16%, respectively., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 Elsevier Ltd. All rights reserved.)

Machine learning algorithms are commonly used for quickly and efficiently counting people from a crowd. Test-time adaptation methods for crowd counting adjust model parameters and employ additional data augmentation to better adapt the model to the specific conditions encountered during testing. The majority of current studies concentrate on unsupervised domain adaptation. These approaches commonly perform hundreds of epochs of training iterations, requiring a sizable number of unannotated data of every new target domain apart from annotated data of the source domain. Unlike these methods, we propose a meta-test-time adaptive crowd counting approach called CrowdTTA, which integrates the concept of test-time adaptation into the meta-learning framework and makes it easier for the counting model to adapt to the unknown test distributions. To facilitate the reliable supervision signal at the pixel level, we introduce uncertainty by inserting the dropout layer into the counting model. The uncertainty is then used to generate valuable pseudo labels, serving as effective supervisory signals for adapting the model. In the context of meta-learning, one image can be regarded as one task for crowd counting. In each iteration, our approach is a dual-level optimization process. In the inner update, we employ a self-supervised consistency loss function to optimize the model so as to simulate the parameters update process that occurs during the test phase. In the outer update, we authentically update the parameters based on the image with ground truth, improving the model's performance and making the pseudo labels more accurate in the next iteration. At test time, the input image is used for adapting the model before testing the image. In comparison to various supervised learning and domain adaptation methods, our results via extensive experiments on diverse datasets showcase the general adaptive capability of our approach across datasets with varying crowd densities and scales.

Background: N7-methylguanosine (m7G) modification has been reported to regulate RNA expression in multiple pathophysiological processes. However, little is known about its role and association with immune microenvironment in heart failure (HF). Results: One hundred twenty-four HF patients and 135 nonfailing donors (NFDs) from six microarray datasets in the gene expression omnibus (GEO) database were included to evaluate the expression profiles of m7G regulators. Results revealed that 14 m7G regulators were differentially expressed in heart tissues from HF patients and NFDs. Furthermore, a five-gene m7G regulator diagnostic signature, NUDT16, NUDT4, CYFIP1, LARP1, and DCP2, which can easily distinguish HF patients and NFDs, was established by cross-combination of three machine learning methods, including best subset regression, regularization techniques, and random forest algorithm. The diagnostic value of five-gene m7G regulator signature was further validated in human samples through quantitative reverse-transcription polymerase chain reaction (qRT-PCR). In addition, consensus clustering algorithms were used to categorize HF patients into distinct molecular subtypes. We identified two distinct m7G subtypes of HF with unique m7G modification pattern, functional enrichment, and immune characteristics. Additionally, two gene subgroups based on m7G subtype-related genes were further discovered. Single-sample gene-set enrichment analysis (ssGSEA) was utilized to assess the alterations of immune microenvironment. Finally, utilizing protein–protein interaction network and weighted gene co-expression network analysis (WGCNA), we identified UQCRC1, NDUFB6, and NDUFA13 as m7G methylation-associated hub genes with significant clinical relevance to cardiac functions. Conclusions: Our study discovered for the first time that m7G RNA modification and immune microenvironment are closely correlated in HF development. A five-gene m7G regulator diagnostic signature for HF (NUDT16, NUDT4, CYFIP1, LARP1, and DCP2) and three m7G methylation-associated hub genes (UQCRC1, NDUFB6, and NDUFA13) were identified, providing new insights into the underlying mechanisms and effective treatments of HF. [ABSTRACT FROM AUTHOR]

This work extends the conventional back-propagation neural network (BPNN) to the classification of Chinese liquors of different flavors according to their Raman spectra. Conformal prediction is applied to assign reliable confidence measures for each classification and support an effective framework to make the machine learning on classification trustable. The BPNN can be used to predict the flavors of Chinese liquors according to their Raman spectra, and a classification rate of 88.96% can be achieved. In order to evaluate each classification, a non-conformity score is defined to generate a P -value for each classification. Moreover, the validity of conformal prediction in online mode is discussed. The number of cumulative errors in the conformal prediction is much less than that without conformal prediction. The relationship between the cumulative error and confidence levels shows that a high confidence level leads to low cumulative errors, but many cumulative errors will occur under a very high confidence level. The result implies that conformal prediction is a useful framework, which can employ classification satisfying a certain level of confidence. Meanwhile, the conformal prediction can improve our classification using a BPNN, when the number of data points is limited. [ABSTRACT FROM AUTHOR]

Quinolone antibiotics have good antibacterial properties and are commonly used antibiotics in the dairy industry. Currently, the problem of excessive antibiotics in dairy products is very serious. As an ultra-sensitive detection technology, Surface-Enhanced Raman Scattering (SERS) was applied to the detection of quinolone antibiotics in this work. In order to classify and quantify three antibiotics (Ciprofloxacin, Norfloxacin, Levofloxacin) with highly similar molecular structures, a combination of magnetic COF-based SERS substrate and machine learning algorithms (PCA-k-NN, PCA-SVM, PCA-Decision Tree) was used. The classification accuracy of the spectral dataset could reach 100% and the results of LOD calculation were: CIP: 5.61 × 10−9M, LEV: 1.44 × 10−8M, NFX: 1.56 × 10−8M. This provides a new method for the detection of antibiotics in dairy products. [Display omitted] • This method realized the detection of three quinolone antibiotics with one SERS substrate. • This method can classify three quinolone antibiotics with highly similar molecular structures with 100% accuracy. • The proposed SERS method has low LODs in milk: CIP: 5.61 × 10−9M, LEV: 1.44 × 10−8M, NFX: 1.56 × 10−8M. [ABSTRACT FROM AUTHOR]

Increasing the accuracy of prediction models in financial markets is an important but difficult task due to the natural complexities of financial time series, which are nonlinear and nonstationary. This challenge has made machine learning methods popular in recent years. However, the noise contained in financial series dramatically distorts the performance of such approaches. This paper proposes an adaptive denoising method called MIC-EMD that is data-driven and removes the noise contained in input features based on the nonlinear relationship between the target variable and the input features. To verify the advantages of MIC-EMD, a simulation experiment is conducted to compare the performance of several representative denoising methods with that of MIC-EMD. Finally, a comprehensive empirical analysis is performed for the trend predictions of six major indexes in Asian markets using three prevalent machine learning methods (SVM, random forest and LightBGM). After the input features are denoised by MIC-EMD, the results reveal the following: (i) its prediction performance outperforms that of the three learning models with input features denoised by state-of-the-art denoising methods such as ICA, WT, WF, EMD, aEMD, P-EMD and S-EMD, e.g., the prediction accuracies of the three machine learning models increase by 9.77%, 13.5% and 12.3%; and (ii) we obtain a prediction accuracy as high as 70%. • The new denoising method is data-driven. • It outperforms prevalent state-of-the-art denoising methods. • It increases prediction accuracies of machine learning methods by 10% or more. • We get accuracies up to 70% of trend predictions for major Asian stock indexes. [ABSTRACT FROM AUTHOR]