Multivariate Sample Research Articles

Multivariate Emulation of Kilometer-Scale Numerical Weather Predictions with Generative Adversarial Networks: A Proof of Concept

Abstract Emulating numerical weather prediction (NWP) model outputs is important to compute large datasets of weather fields in an efficient way. The purpose of the present paper is to investigate the ability of generative adversarial networks (GANs) to emulate distributions of multivariate outputs (10-m wind and 2-m temperature) of a kilometer-scale NWP model. For that purpose, a residual GAN architecture, regularized with spectral normalization, is trained against a kilometer-scale dataset from the AROME Ensemble Prediction System (AROME-EPS). A wide range of metrics is used for quality assessment, including pixelwise and multiscale Earth-mover distances, spectral analysis, and correlation length scales. The use of wavelet-based scattering coefficients as meaningful metrics is also presented. The GAN generates samples with good distribution recovery and good skill in average spectrum reconstruction. Important local weather patterns are reproduced with a high level of detail, while the joint generation of multivariate samples matches the underlying AROME-EPS distribution. The different metrics introduced describe the GAN’s behavior in a complementary manner, highlighting the need to go beyond spectral analysis in generation quality assessment. An ablation study then shows that removing variables from the generation process is globally beneficial, pointing at the GAN limitations to leverage cross-variable correlations. The role of absolute positional bias in the training process is also characterized, explaining both accelerated learning and quality-diversity trade-off in the multivariate emulation. These results open perspectives about the use of GAN to enrich NWP ensemble approaches, provided that the aforementioned positional bias is properly controlled.

OS-DS tracker: Orientation-variant Siamese 3D tracking with Detection based Sampling

LIDAR point clouds are less affected by the weather and possess more depth of field than 2D images. However, sparsity and disorder of point clouds bring challenges for 3D object tracking due to the difficulty in detecting and encoding. In addition, ineffective candidate proposals increase the phenomenon of the loss of target tracking or wrong target tracking. In this paper, we propose an Orientation-variant Siamese 3D object tracking network that utilizes a Detection based Sampling module to generate candidate proposals (OS-DS tracker). The Detection based Sampling module is used to cut the excessive and useless proposals. And the multivariate Gaussian sampling generates the candidate proposals when the objects are not detected. Concretely, we first pre-train PointRCNN Network to globally detect objects from LIDAR point clouds. Then the 3D detected objects are refined by Candidate Regions Sampling to generate candidate proposals. Meanwhile, to make the feature vectors more discriminative, we design an Orientation-variant Siamese Auto-encoder, i.e., tracking loss regress to the intersection over union (IOU) of the template box and the candidate region boxes. Our method is tested on the KITTI dataset and the SHIP Tracking dataset. Compared with the previous state-of-the-arts, the proposed method outperforms in vessel tracking with 5.2%/2.1% improvement in Success and Precision.

Racial disparity and regional variance in healthcare utilization among patients with lung cancer in US hospitals during 2016–2019

BackgroundLung cancer health disparities are related to various patient factors. This study describes regional differences in healthcare utilization and racial characteristics to identify high-risk areas. This study aimed to identify regions and races at greater risk for lung cancer health disparities based on differences in healthcare utilization, measured here by hospital charges and length of stay.MethodsThe National Inpatient Sample of the United States was used to identify patients with lung cancer (n = 92,159, weighted n = 460,795) from 2016 to 2019. We examined the characteristics of the patient sample and the association between the racial and regional variables and healthcare utilization, measured by hospital charges and length of stay. The multivariate sample weighted linear regression model estimated how racial and regional variables are associated with healthcare utilization.ResultsOut of 460,795 patients, 76.4% were white, and 40.2% were from the South. The number of lung cancer patients during the study periods was stable. However, hospital charges were somewhat increased, and the length of stay was decreased during the study period. Sample weighted linear regression results showed that Hispanic & Asian patients were associated with 21.1% and 12.3% higher hospital charges than White patients. Compared with the Northeast, Midwest and South were associated with lower hospital charges, however, the West was associated with higher hospital charges.ConclusionMinority groups and regions are at an increased risk for health inequalities because of differences in healthcare utilization. Further differences in utilization by insurance type may exacerbate the situation for some patients with lung cancer. Hospital managers and policymakers working with these patient populations in identified areas should strive to address these disparities through special prevention programs and targeted financial assistance.

Similarity learning-based fault detection and diagnosis in building HVAC systems with limited labeled data

Machine learning has been widely adopted for fault detection and diagnosis (FDD) in heating, ventilation and air conditioning (HVAC) systems over the past decade due to the ever-increasing availability of massive building operational data. Machine learning-based FDD is flexible and accurate but heavily relies on the availability of sufficient labeled data to develop supervised or unsupervised models. However, collecting labeled data is usually labor-intensive for various types of faulty conditions, significantly limiting the practical implementation of machine learning-based FDD. Therefore, this study proposes a similarity learning-based method using Siamese networks to improve the performance of machine learning-based FDD in applications with limited labeled data. Unlike the conventional supervised approach, the proposed Siamese networks contain two identical long short-term memory subnetworks which take a pair of multivariate time-series samples from the building energy management system as input. The number of training samples can be significantly augmented by generating pairs randomly. In this way, the generalization ability of the machine learning-based FDD is significantly improved in practical applications. Two case studies were designed and conducted using experimental data when labeled data were limited and imbalanced to validate the proposed similarity learning-based method. In case 1, the proposed method improves the fault diagnostic accuracy by at most 45.7% compared with the baseline model when the number of labeled data is limited. In case 2, the proposed method demonstrated better generalization ability when the labeled data is imbalanced.

Convergence Results for Nonlinear Sampling Kantorovich Operators in Modular Spaces

In the present paper, convergence in modular spaces is investigated for a class of nonlinear discrete operators, namely the nonlinear multivariate sampling Kantorovich operators. The convergence results in the Musielak-Orlicz spaces, in the weighted Orlicz spaces, and in the Orlicz spaces follow as particular cases. Even more, spaces of functions equipped by modulars without an integral representation are presented and discussed.

A hierarchical multi-level model for compromise allocation in multivariate stratified sample surveys with non-response problem

Stratified sampling, a widely-applicable probabilistic technique, is especially suited for large and heterogeneously distributed populations. This paper studies optimal allocation of sampling units to minimize variances and cost. Given that not all strata hold equal importance in survey projects, we propose a hierarchical multi-level programming model for compromise allocation in stratified sampling. The model considers multiple objectives across hierarchical levels and addresses the issue of non-response by dividing strata into respondent and non-respondent groups. The survey budget restriction is considered as a constraint in our model. We employ a fuzzy concept-based solution methodology to solve the multi-level allocation problem. A literature-based numerical example describes the model’s applicability and efficiency. Comparative analysis shows our model’s superior efficiency over existing models due to its optimization of allocation across varied strata, non-response consideration, and budget constraint integration. The proposed model promotes flexibility, enhancing representativeness and cost-effectiveness of large-scale surveys, thereby improving decision-making in research and industry settings.

Optimal Stochastic Allocation in Multivariate Stratified Sampling

Optimal Stochastic Allocation in Multivariate Stratified Sampling

Sample size and genetic divergence: a principal component analysis for soybean traits

Biometric techniques, including the principal component analysis, are commonly applied in soybean genetic divergence studies. However, the sample size used in such studies is often determined empirically, neglecting its potential impact on inference interpretation. Therefore, this study aimed (i) to analyze the response of principal components to the number of plants sampled per experimental unit (plot); (ii) to determine the multivariate representative sample size, and (iii) to construct a methodology to predict sample size for principal components as a function of the precision level defined a priori. Six experiments were performed in two locations in Rio Grande do Sul, Brazil, with three experiments being carried out in each location. All experiments were conducted using a complete randomized block design with three repetitions, and 20 soybean genotypes were utilized, resulting in 360 plots (60 plots per experiment). From each plot, twenty plants were sampled, totaling 7200 plants. A resampling bootstrap procedure was applied to the principal component technique for 10 biometric traits. Posteriorly, the sample size was defined based on predefined precision levels, and power logistic models were parametrized to predict the sample size per experimental unit. The precision of the eigenvalues obtained from the principal component analysis gradually improves with larger sample sizes. Eigenvalues that capture a higher variance tend to require smaller sample sizes for accurate estimation. The fitted models demonstrated satisfactory predictive ability in determining the optimal number of plants per experimental unit, serving as a complementary tool for defining sample size at the desired precision level. Eighteen plants per experimental unit are enough to estimate the eigenvalues of the first two soybean principal components reliably. The developed predictive methodology and sample dimensioning per experimental unit can support future research aimed at identifying divergent genotypes, making them useful tools for soybean plant breeding programs.

Multivariate Multiscale Entropy: An Approach to Estimating Vigilance of Driver

Various driver’s vigilance estimation techniques currently exist in the literature. But none of them estimates the driver’s vigilance in the complexity domain. In this research, we propose the recently introduced multivariate multiscale entropy method to fill the above mentioned research gap. We apply this technique to differential entropy features of electroencephalogram and electrooculogram signals to detect driver’s vigilance. Also, we employ it to the percentage of eye closure values to analyse the driver’s cognitive states (awake, tired and drowsy) in the complexity domain. The contribution of this research is to efficiently classify the driver’s cognitive states using a new feature based on multivariate multiscale entropy. The experimental complexity profile curves show the statistically significant differences (p < 0.01) among brain electroencephalogram, forehead electroencephalogram and electrooculogram signals. Moreover, the difference in the multivariate sample entropy across all scales in awake (1.0828 ± 0.4664), tired (0.7841 ± 0.3183) and drowsy (0.2938 ± 0.1664) states are statistically significant (p <0.01). Also, the support vector machine, a machine learning technique, discriminates the driver’s cognitive states with a promising classification accuracy of 76.2%. Therefore, the complexity profile of driver’s cognitive states could be an indicator for vigilance estimation.

Improved multivariate multiscale sample entropy and its application in multi-channel data.

Entropy, as a nonlinear feature in information science, has drawn much attention for time series analysis. Entropy features have been used to measure the complexity behavior of time series. However, traditional entropy methods mainly focus on one-dimensional time series originating from single-channel transducers and are incapable of handling the multidimensional time series from multi-channel transducers. Previously, the multivariate multiscale sample entropy (MMSE) algorithm was introduced for multi-channel data analysis. Although MMSE generalizes multiscale sample entropy and provides a new method for multidimensional data analysis, it lacks necessary theoretical support and has shortcomings, such as missing cross-channel correlation information and having biased estimation results. This paper proposes an improved multivariate multiscale sample entropy (IMMSE) algorithm to overcome these shortcomings. This paper highlights the existing shortcomings in MMSE under the generalized algorithm. The rationality of IMMSE is theoretically proven using probability theory. Simulations and real-world data analysis have shown that IMMSE is capable of effectively extracting cross-channel correlation information and demonstrating robustness in practical applications. Moreover, it provides theoretical support for generalizing single-channel entropy methods to multi-channel situations.

Multivariate sample similarity measure for feature selection with a resemblance model

Feature selection improves the classification performance of machine learning models. It also identifies the important features and eliminates those with little significance. Furthermore, feature selection reduces the dimensionality of training and testing data points. This study proposes a feature selection method that uses a multivariate sample similarity measure. The method selects features with significant contributions using a machine-learning model. The multivariate sample similarity measure is evaluated using the University of California, Irvine heart disease dataset and compared with existing feature selection methods. The multivariate sample similarity measure is evaluated with metrics such as minimum subset selected, accuracy, F1-score, and area under the curve (AUC). The results show that the proposed method is able to diagnose chest pain, thallium scan, and major vessels scanned using X-rays with a high capability to distinguish between healthy and heart disease patients with a 99.6% accuracy.

Multivariate spatial sample reduction of soil chemical attributes by means of application zones

Aim of study: In precision agriculture, the definition of Application Zones (AZs) in agricultural areas consists in delimiting the area in subareas with similar characteristics, using soil chemical attributes. To such end, the use of clustering methods is common. Therefore, the AZs make up a database that can be used to target future soil sampling, thus seeking a possible sample reduction. The objective of this paper is to assess the acquisition of sample configurations, with reduced sample size, contained in application zones generated by spatial multivariate clustering. The sampling protocol proposed in this work evaluated five clustering methods (C-means, Fanny, K-means, Mcquitty, and Ward) for the creation of AZs, and, through these AZs, to obtain reduced sample configurations with 50% and 75% of the initial sampling points. Area of study: Commercial agricultural area, Cascavel, Brazil. Material and methods: Data of the soil chemical attributes from a commercial agricultural area were used, referring to three soybean harvest years (2013-2014; 2014-2015; and 2015-2016). The clustering methods considered a dissimilarity matrix that aggregates the information about the Euclidean distance between the sample elements and the spatial dependence structure of the attributes. Main results: The results indicated division of the agricultural area into two or three AZs for the aforementioned harvest years, considering the K-means method. Comparing all the reduced sample configurations with the initial one, it was observed that the one proportionally reduced by 25% was the most effective to obtain a reduced sample configuration. Research highlights: The sampling protocol using AZs showed that it is possible to reduce the sample size.

Improving ICESat‐2‐based boreal forest height estimation by a multivariate sample quality control approach

Abstract Boreal forest heights are associated with global carbon stocks and energy budgets. The launch of the Advanced Topographic Laser Altimeter System (ATLAS) onboard the NASA's Ice, Cloud and Land Elevation Satellite (ICESat‐2) enables canopy vertical structure measurement at a global scale. However, with a photon‐counting laser system, ICESat‐2 contains high uncertainties in the estimated canopy heights, requiring appropriate quality control before being applied to canopy height modelling. We adopted a multivariate quality control approach (i.e. the Cook's distance) to improve the quality of ICESat‐2 samples. The controlled ICESat‐2 data were then input as the response variable for predicting boreal forest heights based on spatially continuous satellite data and machine learning (ML) regression models. The examined ML regressors include artificial neural networks (ANN), gradient boosting machine (GBM), random forest (RF) and support vector regression (SVR). The proposed quality control effectively removes low‐quality ICESat‐2 samples and enhances the correlations between ICESat‐2 and airborne laser scanning (ALS) observations. Moreover, the controlled ICESat‐2 samples help achieve a trade‐off between sample quality and quantity for all ML regressors, generating close canopy heights to ALS‐derived counterparts. Overall, RF and GBM make better canopy height predictions than ANN and SVR. Of all explanatory variables, the normalized difference vegetation index calculated based on the first red edge band of Sentinel‐2 (NDVIredEdge1) is considered the most important. The proposed quality control on ICESat‐2 sample selection and canopy height model (CHM) development workflow will greatly benefit forest structure investigations in the Arctic community.

Multiobjective intuitionistic fuzzy programming under pessimistic and optimistic applications in multivariate stratified sample allocation problems.

This study investigates the compromise allocation of multivariate stratified sampling with complete response and nonresponse. We have formulated a multivariate stratified sampling problem as a mathematical programming problem to estimate p-population means with complete response and nonresponse for a fixed cost. Then, the compromise allocations for sample designs are determined by implementing intuitionistic fuzzy programming using optimistic and pessimistic solution strategies. A simulation study is carried out using the Stratify R software program to demonstrate the complete solution process. In wildlife, agricultural and marketing-related surveys, the study could be helpful. Also, the national planning policies related to surveys in such cases this study could be helpful. This study is an attempt to solve the sampling optimization problem using the Lingo-18 optimization program.

Abstract 02: Intergenerational Socioeconomic Mobility is Associated With Adulthood All-Cause Mortality in the Hispanic Community Health Study/Study of Latinos

Introduction: Different patterns of socioeconomic mobility (SEM) across the life course may influence the risk of all-cause mortality in adulthood. Lifetime low socioeconomic status and downward social mobility may be associated with a higher risk of mortality. Hypothesis: SEM is associated with mortality, and persons with enduring intergenerational adversity or downward SEM have a higher risk of mortality compared to persons with enduring intergenerational advantage. Methods: This study included 13,459 participants, 18-74 years of age, from the Hispanic Community Health Study/Study of Latinos who participated in the baseline visit (2008-2011) and were not missing key covariate information. Socioeconomic position (SEP) was assessed as parental education in childhood (dichotomized at high school), and a 5-facet index of SEP in adulthood across education, income, employment status, longest held occupation type, and homeownership. The adulthood SEP index was dichotomized at the median value of 2 (range: 0-5). Childhood and adulthood SEP were then combined into a single 4-level measure of SEM (enduring advantage, upward mobility, downward mobility, and enduring adversity). Multivariable sampling weight-adjusted Cox proportional hazards regression was performed to estimate the association between SEM and all-cause mortality. Potential confounders included age, gender, Hispanic/Latino ethnic background, US nativity, years spent in the US, and study center, whereas alcohol and tobacco use, depression score, and BMI were considered potential mediators. Presence of mediation was assessed by calculating total and direct effects via modeling, then evaluating percent change from total to direct effects against a 10% guideline. Results: During the follow-up period of 2008-2019, 515 deaths were recorded. After adjusting for all potential confounders, a significant association was found between adulthood all-cause mortality and SEM overall (F = 7.62, p < 0.0001). Downward mobility had significantly higher hazard of death compared to enduring advantage (aHR: 1.81, 95% CI: 1.17-2.81), although upward mobility and enduring adversity groups were not significant. The direct effect for downward mobility was reduced after adjustment for potential mediators (aHR: 1.49, 95% CI: 0.96, 2.30). Conclusions: Socioeconomic mobility, specifically downward mobility, is associated with all-cause mortality among a diverse cohort of Hispanic/Latino individuals residing in the United States. This association shows evidence of mediation by clinical and behavioral characteristics.

Which Multivariate Multi-Scale Entropy Algorithm Is More Suitable for Analyzing the EEG Characteristics of Mild Cognitive Impairment?

So far, most articles using the multivariate multi-scale entropy algorithm mainly use algorithms to analyze the multivariable signal complexity without clearly describing what characteristics of signals these algorithms measure and what factors affect these algorithms. This paper analyzes six commonly used multivariate multi-scale entropy algorithms from a new perspective. It clarifies for the first time what characteristics of signals these algorithms measure and which factors affect them. It also studies which algorithm is more suitable for analyzing mild cognitive impairment (MCI) electroencephalograph (EEG) signals. The simulation results show that the multivariate multi-scale sample entropy (mvMSE), multivariate multi-scale fuzzy entropy (mvMFE), and refined composite multivariate multi-scale fuzzy entropy (RCmvMFE) algorithms can measure intra- and inter-channel correlation and multivariable signal complexity. In the joint analysis of coupling and complexity, they all decrease with the decrease in signal complexity and coupling strength, highlighting their advantages in processing related multi-channel signals, which is a discovery in the simulation. Among them, the RCmvMFE algorithm can better distinguish different complexity signals and correlations between channels. It also performs well in anti-noise and length analysis of multi-channel data simultaneously. Therefore, we use the RCmvMFE algorithm to analyze EEG signals from twenty subjects (eight control subjects and twelve MCI subjects). The results show that the MCI group had lower entropy than the control group on the short scale and the opposite on the long scale. Moreover, frontal entropy correlates significantly positively with the Montreal Cognitive Assessment score and Auditory Verbal Learning Test delayed recall score on the short scale.

Two-Sample Tests Based on Data Depth

In this paper, we focus on the homogeneity test that evaluates whether two multivariate samples come from the same distribution. This problem arises naturally in various applications, and there are many methods available in the literature. Based on data depth, several tests have been proposed for this problem but they may not be very powerful. In light of the recent development of data depth as an important measure in quality assurance, we propose two new test statistics for the multivariate two-sample homogeneity test. The proposed test statistics have the same χ2(1) asymptotic null distribution. The generalization of the proposed tests into the multivariate multisample situation is discussed as well. Simulations studies demonstrate the superior performance of the proposed tests. The test procedure is illustrated through two real data examples.

Geração de séries sintéticas de vazão a partir de amostragem multivariada de saídas de modelo ARMA contemporâneo

ABSTRACT This work constitutes one of the outcomes of the “Evaluation of hydrological scenario generation models” activity initiated by the Hydrological Scenario Representation Working Group (GT CH) and coordinated by ONS and CCEE. We introduce the LYNX-Series model, a contemporaneous non-periodic and multivariate variation of the autoregressive moving average model (CARMA) for generating synthetic time series of average inflow discharges to reservoirs in the Brazilian National Interconnected System (SIN). Notably, LYNX-Series couples the synthetic series generator with a multivariate sampling process to select a group of synthetic hydrological scenarios based on a similarity criterion with recent historical data. In addition to reducing the computational burden of the hydrothermal dispatch optimization process, the solution aims to enhance the representativeness of synthetic hydrological scenarios. The paper expounds on the theoretical aspects of the model and presents numerical simulations that validate its ability to replicate hydrological behaviors in various Brazilian basins.

Quantifying the Uncertainty in Long-Term Traffic Prediction Based on PI-ConvLSTM Network

This work proposes a novel uncertainty quantification framework for long-term traffic flow prediction (TFP) based on a sequential deep learning model. Quantifying the uncertainty of TFP is crucial for intelligent transportation system (ITS) to make robust traffic congestion analysis and efficient traffic management due to the inherent uncertain and fluctuating nature of traffic flow. However, the performance (e.g., reliability and sharpness) of uncertainty quantification is hard to guarantee, particularly for long-term traffic flow (e.g., one week or two weeks in advance). To this end, this work develops a nonparametric performance-oriented prediction interval (PI) construction approach based on an enhanced sequential convolutional long short-term memory units (ConvLSTM) model, which is named as PI-ConvLSTM. This model can well learn the temporal correlations involved in the multivariate explanatory samples. Specifically, a periodic pattern learning strategy and a performance-oriented loss function are developed to ensure the quality of the derived PIs. Through validating on the real-life England freeway traffic flow dataset, the proposed PI-ConvLSTM proves to be capable of producing the skillful PIs for long-term TFP. For instance, the performance of derived PIs for two-week ahead is 0.175%, 0.198 and 1957.127 in average in terms of reliability, average width and sharpness, respectively. As compared to the benchmark models the proposed model shows at least 68.1% improvement on reliability, 3.4% on average width and 1.7% on sharpness.

Normal but skewed?

SummaryWe propose a multivariate normality test against skew normal distributions using higher‐order log‐likelihood derivatives, which is asymptotically equivalent to the likelihood ratio but only requires estimation under the null. Numerically, it is the supremum of the univariate skewness coefficient test over all linear combinations of the variables. We can simulate its exact finite sample distribution for any multivariate dimension and sample size. Our Monte Carlo exercises confirm its power advantages over alternative approaches. Finally, we apply it to the joint distribution of US city sizes in two consecutive censuses finding that non‐normality is very clearly seen in their growth rates.