Search

With the strong promotion of financial globalization and marketization and the rapid development of financial industry innovation, credit risk management is becoming an important work that banks need to pay more and more attention to. This paper focuses on the construction of the bank credit risk model to carry out research. First of all, according to the credit characteristics selection method, the optimal indicators are combined with the indicators for the correlation test of variables. On this basis, the sample data is calculated using the up-sampling method of SMOTE to complete the algorithm and sampling. After obtaining the data, data preprocessing is carried out using the MinMaxScaler scaler method, and the processed data is inputted into the improved momentum BP neural network to complete the credit risk assessment. Profiling with risk feature data, it is obtained that each feature is not a separate individual from the other. They are interdependent and connected and have obvious correlations. For example, there are 47 features in the dataset with sample missing ratios greater than 0.97, which is too high and indicates an invalid feature. The results of the study show that SMOTE’s up-sampling method and Momentum BP algorithm can quickly utilize big data to provide a more accurate decision basis for bank credit risk assessment.

Numerous approaches targeting hepatic stellate cells (HSCs) have emerged as pivotal therapeutic strategies to mitigate liver fibrosis and are currently undergoing clinical trials. The investigation of herbal drugs or isolated natural active compounds is particularly valuable, due to their multifaceted functions and low risk of side effects. Recent studies have hinted at the potential efficacy of verbascoside (VB) in ameliorating renal and lung fibrosis, yet its impact on hepatic fibrosis remains to be elucidated. This study aims to evaluate the potential effects of VB on liver fibrosis by assessing its ability to inhibit HSC activation. VB demonstrated significant efficacy in suppressing the expression of fibrogenic genes in activated LX-2 cells. Additionally, VB inhibited the migration and proliferation of these activated HSCs by scavenging reactive oxygen species (ROS) and downregulating the AMPK pathway. Furthermore, a biosafe reverse microemulsion loaded with VB (VB-ME) was developed to improve VB's instability and low bioavailability. The optimal formulation of VB-ME was meticulously characterized, revealing substantial enhancements in cellular uptake, ROS-scavenging capacity, and the suppression of HSC activation. [ABSTRACT FROM AUTHOR]

The aim of this study is to solve the problems of the complicated pretreatment and high analytical cost in the detection technology of trace drugs and their metabolites in municipal wastewater. A high-performance magnetic sorbent was fsynthesized for the enrichment of trace drugs and their metabolites in wastewater to develop a magnetic solid-phase extraction pretreatment combined with the acoustic ejection mass spectrometry (AEMS) analytical method. The magnetic nanospheres were successfully prepared by magnetic nanoparticles modified with divinylbenzene and vinylpyrrolidone. The results showed that the linear dynamic range of 17 drugs was 1–500 ng/mL, the recovery was 44–100%, the matrix effect was more than 51%, the quantification limit was 1–2 ng/mL, and the MS measurement was fast. It can be seen that the developed magnetic solid-phase extraction (MSPE) method is a good solution to the problems of the complicated pretreatment and analytical cost in the analysis of drugs in wastewater. The developed magnetic material and acoustic excitation pretreatment coupled with mass spectrometry analysis method can realize the low-cost, efficient enrichment, and fast analysis of different kinds of drug molecules in urban sewage. [ABSTRACT FROM AUTHOR]

As a fundamental component of heatmap-based human pose estimation methods, heatmap decoding is to transform heatmaps into joint coordinates. We found that previous heatmap decoding methods generally ignored the effect of systematic errors introduced by the resolution increaseing operations in the network decoder. This work fills the gap by taking the systematic errors in heatmap decoding into consideration. We proposed a fast method to reduces the systematic and random errors in one shot by error compensation. The proposed method outperforms the previous best method on the COCO and the MPII datasets while being over 2 times faster. Extensive experiments with different networks, resolutions, metrics and datasets have proved the rationality of the proposed idea.

Deep learning techniques have made certain breakthroughs in direction-of-arrival (DOA) estimation in recent years. However, most of the current deep-learning-based DOA estimation methods view the direction finding problem as a grid-based multi-label classification task and require multiple samplings with a uniform linear array (ULA), which leads to grid mismatch issues and difficulty in ensuring accurate DOA estimation with insufficient sampling and in underdetermined scenarios. In order to solve these challenges, we propose a new DOA estimation method based on a deep convolutional generative adversarial network (DCGAN) with a coprime array. By employing virtual interpolation, the difference co-array derived from the coprime array is extended to a virtual ULA with more degrees of freedom (DOFs). Then, combining with the Hermitian and Toeplitz prior knowledge, the covariance matrix is retrieved by the DCGAN. A backtracking method is employed to ensure that the reconstructed covariance matrix has a low-rank characteristic. We performed DOA estimation using the MUSIC algorithm. Simulation results demonstrate that the proposed method can not only distinguish more sources than the number of physical sensors but can also quickly and accurately solve DOA, especially with limited snapshots, which is suitable for fast estimation in mobile agent localization. [ABSTRACT FROM AUTHOR]