Search

Multivariate time-series forecasting (MTSF) plays an important role in diverse real-world applications. To achieve better accuracy in MTSF, time-series patterns in each variable and interrelationship patterns between variables should be considered together. Recently, graph neural networks (GNNs) has gained much attention as they can learn both patterns using a graph. For accurate forecasting through GNN, a well-defined graph is required. However, existing GNNs have limitations in reflecting the spectral similarity and time delay between nodes, and consider all nodes with the same weight when constructing graph. In this paper, we propose a novel graph construction method that solves aforementioned limitations. We first calculate the Fourier transform-based spectral similarity and then update this similarity to reflect the time delay. Then, we weight each node according to the number of edge connections to get the final graph and utilize it to train the GNN model. Through experiments on various datasets, we demonstrated that the proposed method enhanced the performance of GNN-based MTSF models, and the proposed forecasting model achieve of up to 18.1% predictive performance improvement over the state-of-the-art model. [ABSTRACT FROM AUTHOR]

Probiotics have been defined as live microorganisms that are administered in an appropriate amount to provide health benefits to the host animal. In this study, we investigated the effect of L. salivarius DJ-sa-01 secreting the 3D8 single-chain variable fragment (3D8 scFv) on the growth performance, cytokine secretion, and intestinal microbial flora of chickens. The experiment was divided into the control group and L. salivarius expressing 3D8 scFv experimental group. Chicken was fed 109 colonyforming units (CFUs) of wild-type (WT) L. salivarius or 3D8 scFv-secreting L. salivarius daily for 35 days. The administration of L. salivarius expressing 3D8 scFv significantly improved the body weight of chickens compared with the administration of WT L. salivarius. A 16S ribosomal RNA metagenomic analysis showed that Firmicutes, Proteobacteria, Actinobacteria, and Bacteroidetes were the dominant phyla in both experimental groups. At the genus level, Lactobacillus was more abundant (22.82%) in the L. salivarius/3D8 group compared with the WT L. salivarius group. The serum levels of cytokines, such as IL-8, TNF-a, IL-1ß, IFN-?, IL-4, and IGF1, were significantly reduced in the L. salivarius/3D8-treated chickens. In summary, our results suggest that L. salivarius expressing 3D8 scFv could be considered a feed additive for improving the growth performance, immune function, and disease resistance of poultry. [ABSTRACT FROM AUTHOR]

A heat pump with thermal storage system is a system that operates a heat pump during nighttime using inexpensive electricity; during this time, the generated thermal energy is stored in a thermal storage tank. The stored thermal energy is used by the heat pump during daytime. Based on a model of a dual latent thermal storage tank and a heat pump, this study conducts control simulations using both conventional and advanced methods for heating in a building. Conventional methods include the thermal storage priority method and the heat pump priority method, while advanced approaches include the region control method and the dynamic programming method. The heating load required for an office building is identified using TRNSYS (Transient system simulation), used for simulations of various control methods. The thermal storage priority method shows a low coefficient of performance (COP), while the heat pump priority method leads to high electricity costs due to the low use of thermal storage. In contrast, electricity costs are lower for the region control method, which operates using the optimal part load ratio of the heat pump, and for dynamic programming, which operates the system by following the minimum cost path. According to simulation results for the winter season, the electricity costs using the dynamic programming method are 17% and 9% lower than those of the heat pump priority and thermal storage priority methods, respectively. The region control method shows results similar to the dynamic programming method with respect to electricity costs. In conclusion, advanced control methods are proven to have advantages over conventional methods in terms of power consumption and electricity costs. [ABSTRACT FROM AUTHOR]

Background: Korean medicine (KM) has been widely used in Korea. This study aimed to assess the general perceptions of KM, to investigate the patterns of its usage in 2014, and to compare the results with those of an earlier survey from 2011. Methods: A cross-sectional study was conducted with 1000 Korean people. The questionnaire included items regarding trust in KM, reasons for distrust of KM, and visit frequency to KM clinics. This study used methods consistent with those of a 2011 survey to examine changes in attitudes over 3 years. Results: Despite high rates of trust in KM, the visit frequency decreased from 69.3% in 2011 to 63.2% in 2014. Usage among young adults (in their 20s and 30s) was significantly reduced compared to all other age groups. The KM modality most commonly used by participants was acupuncture, whereas the use of moxibustion and cupping therapies has decreased since 2011. Men and women were most likely to distrust KM due to a "lack of scientific evidence" (59.3%) and "suspicion of KM safety" (47.4%), respectively. Conclusions: The findings suggested that KM use and trust in KM were slightly lower in 2014 than in 2011. The decreases were most notable among individuals in their 30s and in the use of moxibustion in KM therapy. This study aimed to produce practical insights by reviewing patterns of KM use and perceptions over time. Additional surveys must be considered to produce a more in-depth analysis. [ABSTRACT FROM AUTHOR]