Search

Abstract To improve the intelligent adjustment ability and energy consumption prediction accuracy of the fresh air system in hospital buildings, this study constructs an energy consumption prediction model based on the Back Propagation Neural Network (BPNN). Meanwhile, it introduces the Genetic Algorithm (GA) and Fuzzy Logic Algorithm (FLA) to optimize the BPNN, thus enhancing the model’s global search ability and robustness. By comparing the proposed optimized model with other models, the study analyzes the advantages of the proposed model in terms of prediction accuracy and convergence speed. Moreover, its practical effectiveness in energy consumption and operational cost optimization is evaluated. The results show that the Genetic Algorithm-Fuzzy Logic Algorithm-Back Propagation (GA-FLA-BP) algorithm performs the best in load prediction, with prediction errors typically below 1.5%, particularly on the 5th and 18th days, demonstrating exceptional performance. Compared to the GA-BP and FLA-BP models, the GA-FLA-BP algorithm exhibits stronger capabilities in handling complex data and uncertainty. Regarding energy consumption and electricity cost optimization, GA-FLA-BP also outperforms other models. Its energy consumption prediction accuracy is 91.5% and an electricity cost prediction accuracy is 90.8%, resulting in savings of 29.2% in energy consumption and 31.2% in costs. Although other algorithms show improvements, GA-FLA-BP remains significantly ahead. Furthermore, the GA-FLA-BP algorithm excels in robustness, consistency, time complexity, and real-time performance. This algorithm demonstrates the highest stability and consistency, the fastest processing speed, and the shortest response time, proving its superior performance in energy consumption management and cost optimization. This study enhances the intelligent adjustment capability of the fresh air system in hospital buildings by optimizing the energy consumption prediction model. Therefore, the study significantly reduces energy consumption and operational costs, improving the efficiency and economy of energy management.

Abstract The transition between yeast and hyphae is crucial for regulating the commensalism and pathogenicity in Candida albicans. The mechanisms that affect the invasion of hyphae in solid media, whose deficiency is more related to the pathogenicity of C. albicans, have not been elucidated. Here, we found that the disruption of VAM6 or VPS41 which are components of the homotypic vacuolar fusion and protein sorting (HOPS) complex, or the Rab GTPase YPT72, all responsible for vacuole fusion, led to defects in hyphal growth in both liquid and solid media, but more pronounced on solid agar. The phenotypes of vac8Δ/Δ and GTR1 OE-vam6Δ/Δ mutants indicated that these deficiencies are mainly caused by the reduced mechanical forces that drive agar and organs penetration, and confirmed that large vacuoles are required for hyphal mechanical penetration. In summary, our study revealed that large vacuoles generated by vacuolar fusion support hyphal penetration and provided a perspective to refocus attention on the role of solid agar in evaluating C. albicans invasion.

Abstract Background The increasingly aging population in current China has encouraged the emergence of the diversified and multi-level elderly care service industry, and the demand for high-quality elderly life with the help of elderly caregivers continues to grow. Methods Based on the existing questionnaire data, this article explores the influencing factors of the treatment level of care staff, and explores their future development prospects. Results The results show that whether they have participated in relevant vocational skills competitions, whether they have worked overtime, whether they have overtime wages, and their monthly income have significant effects on their satisfaction of treatment levels. Elderly care workers who have participated in skills competitions are more satisfied about their salary. In addition, workers who rarely and occasionally work overtime are more satisfied compared with those who have never worked overtime; Caregivers with a monthly income of 5,000–6,999 yuan are more satisfied with their salary and treatment than those with below 3,000 yuan. Conclusion Therefore, in order to better match the supply and demand of care workers, we should provide formal training and skill competitions for them, appropriately increase their salary level and reasonably arrange their working hours, so as to attract more professional talents into elderly care industry.

Image reversible data hiding (RDH) using interpolation technology (IT) provides high single-layer embedding capacity and has become a research hotspot. To solve serious image distortion in existing methods caused by secondary embedding based on interpolated pixels, we propose an RDH method for interpolated images based on modulo operation and prediction-error expansion (PEE). Our method differs from existing RDH methods that only consider embedding data into interpolated pixels. First, an image interpolation method is used to generate an interpolated image, which is partitioned into overlapping $3\times $ 3 blocks. For each block, secret data are embedded into interpolated pixels using the modulo operation and into reference pixel using the PEE. Standard grayscale images of size $512\times512$ were selected from the popular BossBase and BOWS-2 image databases as input images for experiments. Extensive experimental results show that the proposed method avoids pixel values underflow/overflow caused by embedding data and reduces image distortion effectively. For each standard test image, the maximum bits per pixel (BPP) is higher than 1.5, and the corresponding peak signal to noise ratio (PSNR) is higher than 45.0 dB. At the given embedding capacity, the proposed method outperforms recent similar methods regarding the watermarked image quality. Moreover, the proposed method can resist histogram analysis, and the subjective evaluation proves that the proposed method has good imperceptibility.

The loss given default (LGD) is an important credit risk parameter in the regulatory system for financial institutions. Due to the complex structure of the LGD distribution, we propose a new approach, called the hybrid algorithms multi-stage (HMS) model, to construct a multi-stage LGD prediction model and test it on the US Small Business Administration (SBA)’s small business credit dataset. We then compare the model’s performance under four routes by different evaluation metrics. Finally, pertinent business information and macroeconomic features datasets are added for robustness validation. The results show that HMS performs well and stably for predicting LGD, confirming the superiority of the proposed hybrid unsupervised and supervised machine learning algorithm. Financial institutions can apply the approach to make default predictions based on other credit datasets.

Central macrosegregation and coarse microstructure are the main defects in ingot castings of high alloy steels. Rapid cooling associated with continuous casting could be beneficial in improving the internal quality of ingot castings. This work evaluated the effects of superheat and cooling intensity on the solidification structure and macrosegregation in 6Cr13Mo stainless steel billet. Numerical modeling was performed to determine the cooling conditions of a continuous casting billet. The simulated cooling conditions were used to solidify 6Cr13Mo steel samples in a thermal simulator. Experimental results suggest that low superheat of the melt combined with an increased cooling intensity reduce central segregation and form refine microstructure. Relationships between the macrosegregation index, (Km_i), of solute elements in the final solidification zone and secondary dendritic arm spacing (λ2), which inversely proportional to cooling rates, conform to simple monotonic equations. Such monotonic relationships indicate that increasing cooling rates is an effective way of reducing central macrosegregation and refining microstructure of 6Cr13Mo steel billet.

IgA nephropathy (IgAN) is the most common primary glomerulonephritis, characterized by glomerular deposition of IgA immune complexes, mainly produced by B cells under the regulation of CD4+T cells. However, the alterations of specific CD4+T cell subsets and the mechanism of B cells activation in IgAN remain unclear. Therefore, we aimed to investigate the landscape characteristics and role of CD4+T cells in the progression of IgAN. We identified that the proportion of Th2, Th17 and Tfh (follicular helper T) cells in patients with IgAN was significantly higher than that of healthy controls (P < 0.05). Single-cell RNA sequencing of peripheral blood mononuclear cells (PBMCs) showed that Th cells and B cells in patients with IgAN were more activated. Correspondingly, multiplex immunohistochemistry staining of renal biopsy showed increased infiltration of CD4+T and B cells in the kidneys of patients with IgAN. The degree of infiltration was positively correlated with the degree of renal damage. Interestingly, the proportion of Tfh cells in peripheral blood was positively correlated with the severity of proteinuria. Moreover, the proximity position of Tfh cells and B cells suggested that cell-cell interactions between Tfh and B cells were happening in situ. Intercellular communication analysis also showed enhanced interaction between Tfh cells and B cells in IgAN. Our findings suggested that Tfh cells of patients possibly contributed to the progression of IgAN by activating B cells via cell-cell interactions and TNFSF14-TNFRSF14 may be an underlying signaling pathway.

Cognitive tasks engage multiple brain regions. Studying how these regions interact is key to understand the neural bases of cognition. Standard approaches to model the interactions between brain regions rely on univariate statistical dependence. However, newly developed methods can capture multivariate dependence. Multivariate pattern dependence (MVPD) is a powerful and flexible approach that trains and tests multivariate models of the interactions between brain regions using independent data. In this article, we introduce PyMVPD: an open source toolbox for multivariate pattern dependence. The toolbox includes linear regression models and artificial neural network models of the interactions between regions. It is designed to be easily customizable. We demonstrate example applications of PyMVPD using well-studied seed regions such as the fusiform face area (FFA) and the parahippocampal place area (PPA). Next, we compare the performance of different model architectures. Overall, artificial neural networks outperform linear regression. Importantly, the best performing architecture is region-dependent: MVPD subdivides cortex in distinct, contiguous regions whose interaction with FFA and PPA is best captured by different models.