Traffic networks exhibit complex spatial-temporal dependencies, and accurately capturing such dependencies is critical to improving prediction accuracy. Recently, many deep learning models have been proposed for dependency modeling. While numerous developed modeling, most rely on different types of convolutions extract spatial temporal correlations separately. To address this limitation, we propose a novel framework traffic called GraphSAGE-based Dynamic Spatial-Temporal Graph Convolutional...

B-cell epitopes play important roles in vaccine design, clinical diagnosis, and antibody production. Although some models have been developed to predict linear or conformational epitopes, their performance is still unsatisfactory. Hundreds of thousands epitope data accumulated the Immune Epitope Database (IEDB). These can be explored using deep learning methods, order create better predictive for epitopes.After cleaning, we obtained 240,563 peptide samples with experimental evidence from...

Five prenylated chalcones and one allylated chalcone were prepared according to the analysis based on support vector machine (SVM) classification model. Most of synthesized showed potent vasorelaxant activities through evaluation in aortic rings with endothelium pre-contracted by phenylephrine (PE), indicating that experimental good agreement theoretical ones. Structure-activity relationship these compounds substituent pattern number hydroxyl groups crucial for their replacement prenyl group...

A new model for spiral inductors, which covers wide operation frequency range and full design parameters, is proposed by using artificial neural network (ANN). It pointed out that a four-layered superior to three-layered both on the mapping generalization abilities in inductor modeling. For first time, novel physics-based sampling technique adopted modeling procedure. Equipped with this method, ANN achieves better speed accuracy performances though training data are substantially reduced....

As the number of electrical loads in multi-electric aircraft, load characteristics and constraints become increasingly complex, making it difficult to achieve effective management through a single objective. Firstly, typical aircraft distribution network is analysed determine priority relationship management. And then, according power demand characteristics, objective abstracted as multi-objective mixed integer programming problem, system method designed. Finally, implemented using CPLEX OPL...

In view of the need near-space vehicle energy management system to adapt wide input and output voltage range, this study presents a high-efficiency buck–boost converter its control strategy suitable for range. study, three-mode circuit based on analogy is proposed, which realises smooth switching between in boost, buck, through mode. On basis, power regulation unit adopting maximum tracking algorithm developed. Based realisation standardisation conditioning unit, regulator constructed by...

To produce bearings with more reasonable hardness profiles, a finite element model was established to simulate different heat treatment processes. Based on the simulated an artificial neural network fit relationship between process parameters and profiles. Genetic algorithm used optimise profiles confirmation test conducted. The results showed that optimised profile closer optimisation target. mean absolute error of reduced by 68.16% root square 66.72% after optimisation. This method can be...

The output voltage of fuel and solar cells varies extensively is often low. Thus, high step-up DC-DC converters in DC microgrid should be suitable for an extensive input-voltage range application. However, the stress rectifier diodes voltage-source converters. In addition, input current converter continuous to obtain a lifetime cell realize maximum power point tracking (MPPT) PV array. Output use film capacitor, which can improve reliability converter. boost full-bridge integrated proposed....

The two-switch buck-boost converter has many advantages such as step-up and step-down functions, having the same input output polarity, fewer passive components, lower voltage stress of switching devices. So, it is suitable for photovoltaic applications to achieve a wide range conversion. For this reason, paper analyses circuit topology deduces two-mode control converter. Then introduces interleaved strategy current ripple get better performance. Finally, simulation result by using Simulink...

To investigate the effect of hardness distribution on rolling contact fatigue (RCF) properties axlebox bearings under deeper carburizing conditions, a two-dimensional finite element model based continuum damage mechanics was developed. A method combined with an artificial neural network and genetic algorithm used to search longest RCF lifetime within process parameters. The results showed that high surface proper carburized case depth had higher lifetime. 693 HV 1.71 mm were more suitable....

The research of the reliability will be further into substation level with combination main wring and power system.In this paper, system considering effect wiring evaluates by two different methods which are analytical method Monte Carlo simulation method.The is analyzed combined model in order to more accurately calculate it proves necessity on calculation IEEE RTS example systems.

Chromosome straightening plays an important role in karyotype analysis. Common methods usually adopt geometric algorithms, which tend to affect the chromosome banding patterns process of straightening, resulting feature changes, loss details, and poor generalization. To solve these problems, this paper proposes a novel method based on disentanglement representation learning. Our consists two main components: Disentanglement Representation Encoder (DRE) Straightening Generator (SG), where DRE...

In order to solve the problem of accuracy and applicability gas turbine simulation modeling, it fully meets electromechanical transient medium-and long-term dynamic requirements large power grids. The technical principle unit existing mainstream manufacturers was analyzed, route modeling determined, method for whole process proposed. Based on unified model architecture, characteristics different realized through mode selection. includes governor, actuator, prime mover generator. governor...

Abstract The most potent compound in terms of CCR5 antagonistic activity is derivative (IIId).