Predicting the price movement of finance securities like stocks is an important but challenging task, due to uncertainty financial markets. In this paper, we propose a novel approach based on Transformer tackle stock prediction task. Furthermore, present several enhancements for proposed basic Transformer. Firstly, Multi-Scale Gaussian Prior enhance locality Secondly, develop Orthogonal Regularization avoid learning redundant heads in multi-head self-attention mechanism. Thirdly, design...

Compared with single biometric recognition, multimodal recognition based on fingerprint and finger vein has been widely considered because of its convenient sample collection, high security accurate recognition. However, according to our investigation, there is no public dataset collected at the same time. The existing work uses datasets from different sources for research, besides researchers data building their own equipment, which lacks consideration practical applications. This not...

Fuzzy entropy and image thresholding are the most direct effective methods for segmentation. This paper, taking fuzzy Kapur's as optimal objective function, with modified discrete Grey wolf optimizer (GWO) tool, uses pseudotrapezoid-shaped to conduct membership initialization so achieve segmentation finally by means of local information aggregation. Experiment results show that proposed fuzzy-based GWO aggregation algorithm (FMDGWOA) can search out thresholds effectively accurately. In this...

The drive-by bridge health monitoring is to assess the condition using acceleration responses measured on body or axle of instrumented vehicles. vehicle are greatly affected by road surface roughness that makes dynamic information blurred. Instead direct for monitoring, response contact point (CP) between and further explored enhance modal identification. A novel three-step framework proposed extract components related vibration from responses. first step identify input forces two successive...

The computation of image segmentation has become more complicated with the increasing number thresholds, and option application thresholds in thresholding fields have an NP problem at same time. paper puts forward modified discrete grey wolf optimizer algorithm (MDGWO), which improves on optimal solution updating mechanism search agent by weights. Taking Kapur’s entropy as optimized function based discreteness threshold segmentation, firstly discretizes (GWO) then proposes a new attack...

Bridge modal identification using an instrumented test vehicle as a moving sensor is promising but challenging. A key factor to extract bridge dynamic components from responses measured when the operating. new method based on advanced adaptive signal decomposition technique, successive variational mode (SVMD), has been developed estimate parameters of passing vehicle. When bridge-related are extracted decomposition, natural excitation technique and/or random-decrement fitting methods used...

For a vehicle moving over bridge, the vehicle-bridge interaction (VBI) embraces time-varying modal parameters of system. The identification non-stationary characteristics bridge responses due to load is important and remains challenging task. A new method based on improved empirical wavelet transform (EWT) along with ridge detection signals in time-frequency representation (TFR) proposed estimate instantaneous frequencies (IFs) bridge. Numerical studies are conducted using VBI model...

The local scour process around pile foundations under tidal currents exhibits complex nonlinear and nonstationary dynamic characteristics, primarily stemming from the intricate coupling relationship between levels, flow velocity, direction, evolution. In this paper, a novel hybrid machine learning (ML) framework (referred to as GVCBA) is proposed, which consists of grey wolf optimization (GWO), variational mode decomposition (VMD), convolutional neural network (CNN), bidirectional long...

A Markov chain model was introduced to study the crack propagation law of butt welds in U-ribs orthotropic steel decks a long-span box girder suspension bridge, such as Xihoumen Bridge. First, time–frequency characteristics long-term strain monitoring data Bridge were analyzed. wavelet wave analysis used separate stress–time history bridge deck structure caused by environmental temperature. Moreover, eliminate low-amplitude stress variation random excitation and measurement noise,...

The computational complexity grows exponentially for multi-level thresholding (MT) with the increase of number thresholds. Taking Kapur’s entropy as optimized objective function, paper puts forward modified quick artificial bee colony algorithm (MQABC), which employs a new distance strategy neighborhood searches. experimental results show that MQABC can search out optimal thresholds efficiently, precisely, and speedily, are very close to examined by exhaustive In comparison EMO...

Chloride penetration initiates the corrosion of steel reinforcement and causes deterioration an RC bridge in a marine area. This study focuses on probabilistic models chloride process for initiation analysis reinforcement. The model diffusion coefficient is developed by adding correction terms to correct potential bias existing deterministic model. parameters are calibrated selected with Bayesian approach based long-term test data. surface concentration threshold from reviews discussions...

Wi-Fi-based indoor localization has received extensive attention in the academic community. However, most WiFi-based systems have complex models and high delays, which limit universality of these methods. To solve problems, we propose a depthwise separable convolution based passive system (DSCP) using Wi-Fi channel state information (CSI). DSCP is fingerprint-based system, includes an offline training phase online phase. In phase, scenario first divided into different areas to set locations...

Data missing is a common problem in wireless sensor networks. Currently, to ensure the performance of data processing, making imputation for most method before getting into analysis. In this paper, temporal and spatial nearest neighbor values-based (TSNN), new based on values has been presented. First, four have defined from perspective space time dimensions as well geometrical distances, which are bases algorithm that help exploit correlations among nodes with regression tool. Next,...