Weighting Scheme Research Articles

Prediction of PM2.5 Concentration Based on Deep Learning, Multi-Objective Optimization, and Ensemble Forecast

Accurate and stable prediction of atmospheric PM2.5 concentrations is crucial for air pollution prevention and control. Existing studies usually rely on a single model or use a single evaluation criterion in multi-model ensemble weighted forecasts, neglecting the dual needs for accuracy and stability in PM2.5 forecast. In this study, a novel ensemble forecast model is proposed that overcomes these drawbacks by simultaneously taking into account both forecast accuracy and stability. Specifically, four advanced deep learning models—Long Short-Term Memory Network (LSTM), Graph Convolutional Network (GCN), Transformer, and Graph Sample and Aggregation Network (GraphSAGE)—are firstly introduced. And then, two combined models are constructed as predictors, namely LSTM–GCN and Transformer–GraphSAGE. Finally, a combined weighting strategy is adopted to assign weights to these two combined models using a multi-objective optimization algorithm (MOO), so as to carry out more accurate and stable predictions. The experiments are conducted on the dataset from 36 air quality monitoring stations in Beijing, and results show that the proposed model achieves more accurate and stable predictions than other benchmark models. It is hoped that this proposed ensemble forecast model will provide effective support for PM2.5 pollution forecast and early warning in the future.

Employment of the DCC-GARCH Copula Model to explore a link between robotics and artificial intelligence and green crypto investments.

The pursuit of the Sustainable Development Goals (SDGs) requires considerable new green crypto investments. To attract the flow of this investment, it is necessary to develop and apply robotic artificial intelligence (AI) as it has the potential to encourage the adoption of environmental innovation and increase individuals' environmental awareness. Our research employs the DCC-GARCH Copula Model to examine time-varying spillovers and prove interlinkages between the development of AI and green cryptocurrencies in the period from January 1, 2018, to September 8, 2023. Comparing the optimum hedge ratios with the optimal portfolio weights, we demonstrate that the optimal hedge strategy for BOTZ is the most successful one. However, the success of hedging depends on the portfolio's risk profile. Based on our analysis of the cumulative profit profile of different approaches, we continue to believe that the best portfolio weighting strategy is the one that produces positive returns in the middle of 2020 and the first part of 2022 and 2023. This demonstrates that the most profitable diversification approach is not always the most successful one. Our results have important policy implications for investors and governments.

Physics-aware recurrent convolutional neural networks for modeling multiphase compressible flows

Multiphase compressible flow systems can exhibit unsteady and fast-transient dynamics, marked by sharp gradients and discontinuities, and material boundaries that interact with the evolving flow. The transient nature of the dynamics presents challenges to employing artificial intelligence (AI) and data-driven models for predicting flow behaviors. In this study, we explore the potential of physics-aware recurrent convolutional neural networks (PARC) to model the spatiotemporal dynamics of multiphase flows in the presence of shocks and reaction fronts. PARC is a neural network model that incorporates the generic form of the diffusion–advection–reaction equation in its network architecture, which mimics the process of solving the governing equations of fluid flows. In contrast to physics-informed machine learning approaches such as physics-informed neural networks (PINNs) where models are trained to directly minimize the residual of governing equations, PARC takes a dynamical systems viewpoint and does not seek to minimize potentially nonconvex and nonlinear loss terms. To assess the ability of PARC to accurately learn and simulate the physics of multiphase flows, we train and test PARC on various flow simulation problems, including the Burgers’ equation, fluid flow behind a cylindrical cross-section, and unsteady shock interactions with a particle at varying Mach numbers. We analyze PARC’s performance and examine sources of error in its prediction, in terms of differentiation and integration schemes and different weighting strategies for the model update. Based on our observations, we discuss PARC’s capabilities and limitations in multiphase flow applications and propose future research directions.

A Robust CoS-PVNet Pose Estimation Network in Complex Scenarios

Object 6D pose estimation, as a key technology in applications such as augmented reality (AR), virtual reality (VR), robotics, and autonomous driving, requires the prediction of the 3D position and 3D pose of objects robustly from complex scene images. However, complex environmental factors such as occlusion, noise, weak texture, and lighting changes may affect the accuracy and robustness of object 6D pose estimation. We propose a robust CoS-PVNet (complex scenarios pixel-wise voting network) pose estimation network for complex scenes. By adding a pixel-weight layer based on the PVNet network, more accurate pixel point vectors are selected, and dilated convolution and adaptive weighting strategies are used to capture local and global contextual information of the input feature map. At the same time, the perspective-n-point localization algorithm is used to accurately locate 2D key points to solve the pose of 6D objects, and then, the transformation relationship matrix of 6D pose projection is solved. The research results indicate that on the LineMod and Occlusion LineMod datasets, CoS-PVNet has high accuracy and can achieve stable and robust 6D pose estimation even in complex scenes.

SPATIALLY INFORMED INSIGHTS: MODELING PERCENTAGE POVERTY IN EAST JAVA PROVINCE USING SEM WITH SPATIAL WEIGHT VARIATIONS

The East Java Province stands as one of Indonesia's regions grappling with a notably elevated poverty rate, accounting for 11.32% of the populace. A strategic approach employed to comprehend and redress this issue involves the application of spatial analysis, wherein spatial factors are intricately integrated into the modeling and cartographic representation of poverty data. The primary objective of this research is to discern the principal determinants influencing the incidence of poverty in East Java Province, employing data reflective of the population's poverty percentages within the province for the year 2021. The study incorporates six pivotal variables, namely: the population poverty rate, open unemployment rate, labor force participation rate, average years of schooling, adjusted per capita expenditure, and the gross regional domestic product (GRDP), predicated on adjusted expenditure. Diverse weighting schemes are applied based on both distance (1) and contiguity (2). The optimal predictive model utilized is the Spatial Error Model (SEM) incorporating a Distance Band Weighing (DBW) mechanism with a designated maximum distance ( ) of 75000 meters. Outcomes indicate that the variable wielding the most substantial influence on the poverty percentage in East Java Province is the average years of schooling. Specifically, an increase in the pursuit of formal education manifests as a negative correlate to the poverty percentage, implying an inverse relationship. Moreover, the SEM model adheres to the requisite assumptions, encompassing (1) the normality of residuals, (2) homogeneity of residuals, and (3) non-spatial autocorrelation of residuals. Comparative analyses reveal that the SEM model utilizing DBW yields diminished values for MAE, MSE, RMSE, AIC, and MAPE in comparison to its linear regression counterpart. Furthermore, the pseudo- values obtained from the SEM surpass those derived from the linear regression model. Rigorous likelihood ratio tests underscore substantial disparities between the SEM and linear regression models, with the former proving more efficient and markedly enhancing the model's explanatory prowess concerning variations in the dataset.

Measuring digital development: ranking using data envelopment analysis (DEA) and network readiness index (NRI)

The Network Readiness Index (NRI) is one of the most prominent indicators that shows the digital development of countries. In contrast to the International Digital Economy and Social Index (I-DESI) of the European Union (EU), in 2020, it showed the development of 134 countries compared to 45 countries in I-DESI of EU, which measures only the most developed countries. The aim of this paper is to provide a viable alternative framework to the equal weights scheme of the original NRI scoring model using the Data Envelopment Analysis (DEA) Without Explicit Input (WEI) method and Common Weight Analysis (CWA) method. After determining the common weights, we compare the digital development of the countries in the NRI dataset based on the results obtained, focusing on the countries of the Central and Eastern European (CEE) region and the former Soviet Union.

Portfolio Validity Test in the Chinese Market

Modern portfolio theory is one of the most classic theories in financial investment theory. It was first released by a famous economist, Markowitz. As the most important critical model in investment science, it provides investors with scientific investment allocation methods, which are favored by many investors. It has been widely used and researched in the world. Modern portfolio theory analyses the historical returns of investment objects and guides future investment behavior. However, it is still controversial whether this theory is reasonable and adequate in Chinas stock market. China's stock market is more susceptible to fluctuations than those in Europe and the US, and many of these factors stem from China's strategic policies in various areas. Since the end of the epidemic, China's economic development has also slowed down. In this context, this paper will construct the maximum Sharpe ratio strategy, market capitalization strategy, and average weight strategy. Through the performance evaluation indexes of these three strategies' performance comparison, this paper will verify whether the portfolio strategy is real and effective in China's stock market and can bring more than significant market returns to investors.

Development of Divergence and Interdependence-based Hybrid Weighting Scheme (DIHWS) for accurate assessment of regional drought

Development of Divergence and Interdependence-based Hybrid Weighting Scheme (DIHWS) for accurate assessment of regional drought

Robust edge-preserving image smoothing based on complementary weighting scheme

Robust edge-preserving image smoothing based on complementary weighting scheme

Load Prediction in Double-Channel Residual Self-Attention Temporal Convolutional Network with Weight Adaptive Updating in Cloud Computing.

When resource demand increases and decreases rapidly, container clusters in the cloud environment need to respond to the number of containers in a timely manner to ensure service quality. Resource load prediction is a prominent challenge issue with the widespread adoption of cloud computing. A novel cloud computing load prediction method has been proposed, the Double-channel residual Self-attention Temporal convolutional Network with Weight adaptive updating (DSTNW), in order to make the response of the container cluster more rapid and accurate. A Double-channel Temporal Convolution Network model (DTN) has been developed to capture long-term sequence dependencies and enhance feature extraction capabilities when the model handles long load sequences. Double-channel dilated causal convolution has been adopted to replace the single-channel dilated causal convolution in the DTN. A residual temporal self-attention mechanism (SM) has been proposed to improve the performance of the network and focus on features with significant contributions from the DTN. DTN and SM jointly constitute a dual-channel residual self-attention temporal convolutional network (DSTN). In addition, by evaluating the accuracy aspects of single and stacked DSTNs, an adaptive weight strategy has been proposed to assign corresponding weights for the single and stacked DSTNs, respectively. The experimental results highlight that the developed method has outstanding prediction performance for cloud computing in comparison with some state-of-the-art methods. The proposed method achieved an average improvement of 24.16% and 30.48% on the Container dataset and Google dataset, respectively.

Forecast combination in agricultural economics: Past, present, and the future

AbstractForecasts are common in agricultural settings where they are routinely used for decision‐making. The advent of the computer age has allowed for rapid generation of individual forecasts that can be updated in real time. It is well known that the selection and use of a single forecast can expose the forecaster to serious error as a result of model mis‐specification. Forecast combination avoids this problem by combining information from different forecasts. Although forecast combination can be as simple as averaging across forecasts, advances in machine learning have made it possible to combine forecasts according to more complicated weighting schemes and criteria. We provide an overview of forecast combination techniques, including those at the frontier of current practice and involving machine learning. We also provide a retrospective on the use of forecast combination in agricultural economics and prospects for the future. Several of the techniques are illustrated in an application to forecasting nationwide corn and soybean planted acreage and we demonstrate how forecast combination can improve expert USDA projections.

Light Weight Concurrent Scheduling of Mobile Sink routing protocol towards Minimization of Propagation Delay against Spatial and Temporal Uncertainties of Cluster based Wireless Sensor Network

Wireless Sensor Network highly exposed to certain security attacks due to various uncertainties of the network which lead to node compromise. Hence considerable attention has been made by researcher during scheduling of the mobile sink to data collection of sensed information from the sensor nodes. Many secure data aggregation mechanism has been proposed to withstand the network against these kind of attack. Despite of many advantageous, those architectures will lead to several limitations such as large propagation delay and high energy consumption and spatial and temporal uncertainties. In order to mitigate those challenges, Light Weight Concurrent Scheduling of Mobile Sink routing protocol towards Minimization of Propagation Delay against Spatial and Temporal Uncertainties of Wireless Sensor Network is designed. It should be capable achieving energy efficiency and collision avoidance due to uncertainty in transmission and reception. Proposed model eliminates the packet collision and poor channel utilization challenges. Initially, node is partitioned and cluster head is selected based on the highest energy density of the node in the particular location. Employing Full duplex channel easily examines the spatial and temporal characteristic of the node through utilization cache information of routing table. It further helps to build the light weight scheme with behavioral strategies of the nodes for secure propagation of the sensed data to the mobile sink. Light weight scheme deployed in mobile sinks estimate the node trustworthiness and simultaneously collects the secure aggregated data packets of the sensor node through cluster head against dynamic time slots concurrently .Comparing with traditional secure routing architecture, simulation results demonstrates that proposed architecture can reduce the propagation delay against various uncertainties and maximizes the life time of the network on increasing the networking performances with respect to packet delivery ratio, throughput and propagation delay. Finally proposed model obtains the cooperative communication of sensor node information to the base station through mobile sink on heterogeneous wireless sensor network environment. DOI: https://doi.org/10.52783/pst.387

Concept drift detection methods based on different weighting strategies

Concept drift detection methods based on different weighting strategies

Alternative weighting schemes for fine‐tuned extended similarity indices

AbstractExtended similarity indices (i.e., generalization of pairwise similarity) have recently gained importance because of their simplicity, fast computation, and superiority in tasks like diversity picking. However, they operate with several meta parameters that should be optimized. Earlier, we extended the binary similarity indices to “discrete non‐binary” and “continuous” data; now we continue with introducing and comparing multiple weighting functions. As a case study, the similarity of CYP enzyme inhibitors (4016 molecules after curation) was characterized by their extended similarities, based on 2D descriptors, MACCS and Morgan fingerprints. A statistical workflow based on sum of ranking differences (SRD) and analysis of variance (ANOVA) was used for finding the optimal weight function(s). Overall, the best weighting function is the fraction (“frac”), which corresponds to the principle of parsimony. Optimal extended similarity indices were also found, and their differences are revealed across different data sets. We intend this work to be a guideline for users of extended similarity indices regarding the various weighting options available. Source code for the calculations is available at https://github.com/mqcomplab/MultipleComparisons.

Quantitative modelling of type Ia supernovae spectral time series: Constraining the explosion physics

Abstract Multiple explosion mechanisms have been proposed to explain type Ia supernovae (SNe Ia). Empirical modelling tools have also been developed that allow for fast, customised modelling of individual SNe and direct comparisons between observations and explosion model predictions. Such tools have provided useful insights, but the subjective nature with which empirical modelling is performed makes it difficult to obtain robust constraints on the explosion physics or expand studies to large populations of objects. Machine learning accelerated tools have therefore begun to gain traction. In this paper, we present riddler, a framework for automated fitting of SNe Ia spectral sequences up to shortly after maximum light. We train a series of neural networks on realistic ejecta profiles predicted by the W7 and N100 explosion models to emulate full radiative transfer simulations and apply nested sampling to determine the best-fitting model parameters for multiple spectra of a given SN simultaneously. We show that riddler is able to accurately recover the parameters of input spectra and use it to fit observations of two well-studied SNe Ia. We also investigate the impact of different weighting schemes when performing quantitative spectral fitting and show that best-fitting models and parameters are highly dependent on the assumed weighting schemes and priors. As spectroscopic samples of SNe Ia continue to grow, automated spectral fitting tools such as riddler will become increasingly important to maximise the physical constraints that can be gained in a quantitative and consistent manner.

Distributed Physics-Informed machine learning strategies for two-phase flows

This study delves into applying Distributed Learning Machines (DLM), a subset of Physics-Informed Neural Networks (PINN), in tackling benchmark challenges within two-phase flows. Specifically, two variants of DLM, namely Distributed Physics-Informed Neural Networks (DPINN) and Transfer Physics-Informed Neural Networks (TPINN), are studied. The DLM architecture strategically divides the global domain into distinct non-overlapping sub-domains, with interconnected solutions facilitated by interface conditions embedded in the loss function. Forward and inverse benchmark problems in two-phase flows are explored: (a) bubble in a reversing vortex and (b) Bubble Rising under Buoyancy. The Volume of Fluid (VOF) method handles interface transport in both scenarios, with the inverse problem incorporating interface-position data during the training phase. The forward problem highlights the effectiveness of DPINN in capturing the interface using a simple transport equation. The distinctive contribution of this work lies in its exploration of the inverse problem, offering insights into the scalability of distributed architectures when dealing with a system of governing equations. Following the validation of an initial PINN model against Computational Fluid Dynamics (CFD) data, the study extends to DPINN and TPINN. A parametric study optimizes network hyperparameters, emphasizing the regularization of loss terms within the DPINN loss function. A self-adaptive weighting strategy based on a Gaussian probabilistic model dynamically adjusts loss weights during training to overcome challenges associated with manual parameter tuning. The evaluation of accuracy against CFD data and published results underscore the efficacy of DLMs in addressing two-phase flow problems. Additionally, the computational efficiency of distributed networks is explored compared to traditional PINNs.

An Empirical Study On Imaging Sensitivity Of SKA1-LOW Based on Point Source Sensitivity

The Square Kilometre Array (SKA) radio telescope will achieve unparalleled sensitivity and angular resolution, substantially progressing our research into the formation and development of the early universe. Sensitivity is one of the most critical telescope metrics, often used to determine the observing mode and observing time of a radio source. It is also used to determine imaging parameters to balance imaging sensitivity and spatial resolution during data processing. In this paper, to meet the needs of continuum imaging pipeline data processing for the SKA1-LOW telescope, we derive equations for point-source-sensitivity (PSS) calculations under multichannel, multipolarization, discrete frequency sampling, discrete time sampling, and different weighting schemes. After validating the correctness of the equation, we improved and refined the sensitivity calculation application in the Radio Astronomy Simulation, Calibration, and Imaging Library, and further calculated the variation of the PSS versus angular resolution for the full-scale SKA1-LOW with different weighting schemes, and provided SKA1-LOW broadband image performance as functions of angular scale for four frequencies. These results can provide not only a theoretical basis for the construction and commissioning of the SKA1-LOW telescope but also guidance for imaging processing in future scientific data processing.

Mitigating the effect of source structure in geodetic VLBI by re-weighting observations using closure delays and baseline-to-jet orientation

An ideal target for geodetic very long baseline interferometry (VLBI) is a strong and point-like radio source. In reality, most celestial sources used in geodetic VLBI have spatial structure. This is as a major source of error in VLBI Global Observing System (VGOS) and also affects legacy S/X observations. Source structure causes a systematic delay, which can affect the geodetic estimates if not modelled or otherwise accounted for. In this work, we aim to mitigate its impact by extending the stochastic model used in the least-squares fitting of the VLBI group delays. We have developed a weighting scheme to re-weight the observations by parameterizing the source structure component in terms of closure delays and jet orientation relative to the observing baseline. It was implemented in the Vienna VLBI Software. To assess the performance of the extended stochastic model, we analysed the CONT17 legacy sessions and generated suitable reference solutions for comparison. The effects of re-weighting were evaluated with respect to the session fit statistics, source-wise residuals, and geodetic parameters. We find that this relatively simple noise model consistently improves the session fit by about 5% with moderate variation from session to session. The geodetic estimates are not affected to a significant level by this new weighting method. Source-wise we see improved post-fit residuals for 63 out of a total of 91 sources observed.

DeepOTF: Learning Equations-constrained Prediction for Electromagnetic Behavior

High-quality passive devices are becoming increasingly important for the development of mobile devices and telecommunications, but obtaining such devices through simulation and analysis of electromagnetic (EM) behavior is time-consuming. To address this challenge, artificial neural network (ANN) models have emerged as an effective tool for modeling EM behavior, with NeuroTF being a representative example. However, these models are limited by the specific form of the transfer function, leading to discontinuity issues and high sensitivities. Moreover, previous methods have overlooked the physical relationship between distributed parameters, resulting in unacceptable numeric errors in the conversion results. To overcome these limitations, we propose two different neural network architectures: DeepOTF and ComplexTF. DeepOTF is a data-driven deep operator network for automatically learning feasible transfer functions for different geometric parameters. ComplexTF utilizes complex-valued neural networks to fit feasible transfer functions for different geometric parameters in the complex domain while maintaining causality and passivity. Our approach also employs an Equations-constraint Learning scheme to ensure the strict consistency of predictions and a dynamic weighting strategy to balance optimization objectives. The experimental results demonstrate that our framework shows superior performance than baseline methods, achieving up to 1700 × higher accuracy.

Improving a methodology for estimating the cross-country ability of all-wheel-drive vehicles

The object of this study is the cross-country ability of four-wheel drive vehicles under off-road conditions. Based on the results of analysis of known studies, it was determined that the procedures of experimental assessment of the load-bearing capacity of bearing surfaces (BS), based on the use of the cone index, are used in NATO member countries. They take into account the characteristics of the surface of the theater of operations and make it possible to determine the approximate speeds of all-wheel drive vehicles with the help of standardized computer software. In the procedures for assessing the load-bearing capacity of BSs, which use the soil deformation module E, there is no possibility for calculating the speed of the car with known axle loads. In addition, they provide for the determination of a number of parameters of the bearing surface, such as the coefficient of adhesion, the angle of internal friction in the soil, and the shear modulus during the formation of a track. This makes them much more difficult compared to the procedures used by NATO. Therefore, the problem that is solved in this work is to improve the methodology for assessing the reference cross-country ability of four-wheel drive vehicles. Based on the results of scientific research, motion modeling was carried out in the MATLAB Simulink environment to determine parameters of reference cross-country ability of four-wheel drive vehicles. In order to determine the cross-country ability of samples of wheeled military vehicles (WMV) based on the indicator of the maximum pressure value (mean measure pressure – MMP), the conducted experimental studies were analyzed. They showed a difference in the speed modes of vehicles of the same gross weight but different layout schemes, within 11–12 %. Therefore, the methodology for assessing the cross-country ability based on MMP has been improved to take into account the features of the layout of WMV samples. Specifically, taking into account the different values of loads on the front axle of vehicles with hoodless and hood layouts at the same gross weight. In general, the improved procedure makes it possible to give a quantitative assessment of the movement mobility of WMV samples with various layout schemes and an assessment of their potential cross-country ability on the basis of the proposed refinement of MMP calculation