Multivariate Data Research Articles

Predicting competitive alpine skiing performance by multivariable statistics—the need for individual profiling

IntroductionPredicting competitive alpine skiing performance using conventional statistical methods has proven challenging. Many studies assessing the relationship between physiological performance and skiing outcomes have employed statistical methods of questionable validity. Furthermore, the reliance on Fédération Internationale de Ski (FIS) points as a performance outcome variable presents additional limitations due to its potential unreliability in reflecting short-term, sport-specific performance. These factors complicate the selection of appropriate tests and the accurate prediction of competitive outcomes.MethodThis study aimed to evaluate the predictive power of a generalized physiological test battery for alpine skiing performance, as measured by FIS points, utilizing multivariable data analysis (MVDA). Physiological test results from a total of twelve (n = 12) world-class female skiers were included in the analysis.ResultsThe result on goodness of regression (R2) and goodness of prediction (Q2) in this study indicate that valid Orthogonal Projection to Latent Structures (OPLS) models for both Slalom and Giant Slalom can be generated (R2 = 0.39 to 0.40, Q2 = 0.21 to 0.15), but also that competition performance still cannot be predicted at a group level (low Q2). In contrast, higher predictive power of competitive performance was achieved on an individual level using the same data (R2 = 0.88 to 0.99 and Q2 = 0.64 to 0.96).DiscussionThe findings of this investigation indicate that the selected tests employed in this study exhibit limited generalizability for the assessment of elite alpine skiers, as the predictive value of specific physiological parameters on competitive performance appears to be highly athlete-dependent.

Causal analysis for multivariate integrated clinical and environmental exposures data

Causal analysis for multivariate integrated clinical and environmental exposures data

Persistent, Private and Mobile genes: a model for gene dynamics in evolving pangenomes.

The pangenome of a species is the set of all genes carried by at least one member of the species. In bacteria, pangenomes can be much larger than the set of genes carried by a single organism. Many questions remain unanswered regarding the evolutionary forces shaping the patterns of presence/absence of genes in pangenomes of a given species. We introduce a new model for bacterial pangenome evolution along a species phylogeny that explicitly describes the timing of appearance of each gene in the species and accounts for three generic types of gene evolutionary dynamics: persistent genes that are present in the ancestral genome, private genes that are specific to a given clade, and mobile genes that are imported once into the gene pool and then undergo frequent horizontal gene transfers. We call this model the Persistent-Private-Mobile (PPM) model. We develop an algorithm fitting the PPM model and apply it to a dataset of 902 Salmonella enterica genomes. We show that the best fitting model is able to reproduce the global pattern of some multivariate statistics like the gene frequency spectrum and the parsimony vs. frequency plot. Moreover, the gene classification induced by the PPM model allows us to study the position of accessory genes on the chromosome depending on their category, as well as the gene functions that are most present in each category. This work paves the way for a mechanistic understanding of pangenome evolution, and the PPM model developed here could be used for dynamics-aware gene classification.

Revisiting the Metabolism of Donepezil in Rats Using Non-Targeted Metabolomics and Molecular Networking

Background/Objectives: Although donepezil, a reversible acetylcholinesterase inhibitor, has been in use since 1996, its metabolic characteristics remain poorly characterized. Therefore, this study aims to investigate the in vivo metabolism of donepezil using liquid chromatography–high-resolution mass spectrometry (LC-HRMS) based on a molecular networking (MN) approach integrated with a non-targeted metabolomics approach. Methods: After the oral administration of donepezil (30 mg/kg) in rats, urine, feces, and liver samples were collected for LC-HRMS analysis. Chromatographic and spectrometric data were processed through MN and multivariate data analysis to identify the in vivo metabolites of donepezil. Results: A total of 50 metabolites were characterized, including 23 newly identified metabolites. Donepezil was biotransformed by O-demethylation, N-debenzylation, and hydroxylation, and these metabolites are further conjugated with glucuronic acid and sulfurous acid. N-Desbenzyldonepezil (M4), didesmethyldonepezil (M5), and N-desbenzyldonepezil (M4) were identified as the most abundant metabolites in urine, feces, and liver samples, respectively. Conclusions: The metabolic characteristics of donepezil in rats were comparable to those in humans, indicating that a rat is a reliable model for studying donepezil metabolism. This study indicates that a MN approach combined with a metabolomics approach is a reliable tool to identify unknown metabolites of drugs and drug candidates.

Predictors of Consumer Intention Towards Upcycling Practice and Consumption in Brazil

Objective: The aim of this study is to investigate the predictors that influence consumer intention to adopt upcycling practices in Brazil. Theoretical Framework: Upcycling emerges as an innovative consumption approach, standing out as a more sustainable alternative compared to traditional methods of goods production. Method: The research adopted a quantitative methodology, using a survey to assess consumers' perceptions of upcycling. Multivariate data analysis and structural equation modeling were applied to identify correlations between the proposed constructs and gain insights into participants' perspectives. Results and Discussion: The results indicate that the factors positively predicting consumer willingness to engage in upcycling are, in order of influence: behavioral intention, followed by the role of beliefs and perceived facilitating conditions, which include skills, knowledge, and inspiration. Research Implications: The research makes a theoretical contribution by providing valuable insights, derived from primary data, into consumers' perceptions of their willingness to engage in upcycling. Practically, the study offers significant implications for management and public policies by promoting a more sustainable consumption approach, raising environmental awareness, and generating tangible social benefits. Originality/Value: The study advances the development of the proposed theoretical model by including and analyzing the influence of predictive factors such as autonomy, empowerment, and entrepreneurial opportunities in the creation of new businesses. The results demonstrate that these factors play crucial roles and make significant contributions to the advancement of consumer behavior theory, deepening the understanding of this phenomenon in practice and promoting more sustainable consumption practices.

Lost Highways: An Examination of the Question of Risk Involved in Sexual Homicides of Hitchhiking Victims.

Despite cultural references to the dangers of hitchhiking, particularly for sexual homicide, no published research investigates these incidents from both an offender and crime scene perspective. Using the Sexual Homicide International Database (SHIelD), we explore lifestyle risk by comparing sexual homicide cases involving hitchhiking victims to those involving victims engaged in sex trade work. The results, based on the use of bivariate and multivariate statistics, indicate that offenders view hitchhiking victims as opportunities for confinement without physical restraint, often engaging in sexual acts and theft. While not primarily sadistic or sexually deviant, many offenders partake in criminal activities, exhibit psychological disorders, and possess weapons. Hitchhiking facilitates perpetrator-victim encounters due to its environmental characteristics. Victims in the sex trade, typically found in isolated locations, are at the mercy of offenders who drive them to unknown destinations. In contrast, murderers targeting low-risk victims display more sexual preoccupations, inserting foreign objects and engaging in postmortem activities. These distinctions suggest distinct offender profiles for each lifestyle.

MetaboLabPy—An Open-Source Software Package for Metabolomics NMR Data Processing and Metabolic Tracer Data Analysis

Introduction: NMR spectroscopy is a powerful technique for studying metabolism, either in metabolomics settings or through tracing with stable isotope-enriched metabolic precursors. MetaboLabPy (version 0.9.66) is a free and open-source software package used to process 1D- and 2D-NMR spectra. The software implements a complete workflow for NMR data pre-processing to prepare a series of 1D-NMR spectra for multi-variate statistical data analysis. This includes a choice of algorithms for automated phase correction, segmental alignment, spectral scaling, variance stabilisation, export to various software platforms, and analysis of metabolic tracing data. The software has an integrated help system with tutorials that demonstrate standard workflows and explain the capabilities of MetaboLabPy. Materials and Methods: The software is implemented in Python and uses numerous Python toolboxes, such as numpy, scipy, pandas, etc. The software is implemented in three different packages: metabolabpy, qtmetabolabpy, and metabolabpytools. The metabolabpy package contains classes to handle NMR data and all the numerical routines necessary to process and pre-process 1D NMR data and perform multiplet analysis on 2D-1H, 13C HSQC NMR data. The qtmetabolabpy package contains routines related to the graphical user interface. Results: PySide6 is used to produce a modern and user-friendly graphical user interface. The metabolabpytools package contains routines which are not specific to just handling NMR data, for example, routines to derive isotopomer distributions from the combination of NMR multiplet and GC-MS data. A deep-learning approach for the latter is currently under development. MetaboLabPy is available via the Python Package Index or via GitHub.

Adapting to the stream: an instance-attention GNN method for irregular multivariate time series data

Multivariate time series (MTS) data are vital for various applications, particularly in machine learning tasks. However, challenges such as sensor failures can result in irregular and misaligned data with missing values, thereby complicating their analysis. While recent advancements use graph neural networks (GNNs) to manage these Irregular Multivariate Time Series (IMTS) data, they generally require a reliable graph structure, either pre-existing or inferred from adequate data to properly capture node correlations. This poses a challenge in applications where IMTS data are often streamed and waiting for future data to estimate a suitable graph structure becomes impractical. To overcome this, we introduce a dynamic GNN model suited for streaming characteristics of IMTS data, incorporating an instance-attention mechanism that dynamically learns and updates graph edge weights for realtime analysis. We also tailor strategies for high-frequency and low-frequency data to enhance prediction accuracy. Empirical results on real-world datasets demonstrate the superiority of our proposed model in both classification and imputation tasks.

SOM modelling and prospectivity mapping of surface geochemical data in Au and multi-metal mineral exploration: example from northern Finland

Regional till and weathered bedrock geochemical data sets, provide a basis for data analysis and modelling in the glaciated terrains. These large-scale surface geochemical data sets have great potential in mineral exploration, especially when machine learning and clustering methods are used to reduce the dimensionality of multivariate data sets. Here, self-organising maps (SOM) followed by the k-means clustering were used to create SOM of the target areas for initial modelling and prospectivity mapping of Au prospecting in central Lapland, northern Finland. Due to the nature of the till and weathered bedrock data being legacy data, levelling effort was made between the map sheets. The targeting till geochemical data set did not contain the typical indicator elements for gold. Instead, elements associated with the gold deposits, within the study area, were used. Indicator associations in this study were Ni-Co with possible Cu. Resulting elemental clusters from SOM were assigned as interesting clusters according to their distribution of elements. For the till, two potential clusters were deemed, Ni-Co-Cr and Cu-V-Co. For the weathered bedrock, three clusters were specified, Ni-Co-Cr, V-Cu, and Cu-Co. This study shows the potential of using legacy data sets for early targeting stages of mineral exploration, potentially reducing the footprint of mineral exploration.

Quadratic Forms in Random Matrices with Applications in Spectrum Sensing

Quadratic forms with random kernel matrices are ubiquitous in applications of multivariate statistics, ranging from signal processing to time series analysis, biomedical systems design, wireless communications performance analysis, and other fields. Their statistical characterization is crucial to both design guideline formulation and efficient computation of performance indices. To this end, random matrix theory can be successfully exploited. In particular, recent advancements in spectral characterization of finite-dimensional random matrices from the so-called polynomial ensembles allow for the analysis of several scenarios of interest in wireless communications and signal processing. In this work, we focus on the characterization of quadratic forms in unit-norm vectors, with unitarily invariant random kernel matrices, and we also provide some approximate but numerically accurate results concerning a non-unitarily invariant kernel matrix. Simulations are run with reference to a peculiar application scenario, the so-called spectrum sensing for wireless communications. Closed-form expressions for the moment generating function of the quadratic forms of interest are provided; this will pave the way to an analytical performance analysis of some spectrum sensing schemes, and will potentially assist in the rate analysis of some multi-antenna systems.

A conceptual model for multivariate data integration in reservoir modeling: exploring the role of neural networks in petrophysical analysis

Reservoir modeling is crucial for optimizing oil and gas production, as it helps characterize subsurface properties like porosity, permeability, and fluid saturation. Integrating diverse data sources such as well logs, seismic data, and core analyses presents a significant challenge due to their differing scales, resolutions, and data types. Traditional methods often struggle to accurately integrate these data sources, resulting in suboptimal predictive accuracy. This review proposes a conceptual model that leverages neural networks to integrate multivariate data for enhanced reservoir characterization. Neural networks, with their ability to handle nonlinear relationships and large, complex datasets, offer a transformative approach to unify diverse data inputs, resulting in more accurate predictions of reservoir properties. The model focuses on using feedforward neural networks (FNNs), convolutional neural networks (CNNs), and autoencoders to merge well logs, seismic interpretations, and core sample data. This integration aims to improve the identification of productive zones, reservoir boundaries, and fluid distributions, thereby enhancing reservoir models. Additionally, the review explores how neural networks can manage data gaps, predict missing values, and quantify uncertainty, which are common challenges in reservoir studies. By automating data processing and facilitating real-time analysis, neural networks can accelerate decision-making in reservoir management and field development. This framework highlights the potential of AI-driven techniques to revolutionize reservoir modeling, offering a pathway to more efficient, accurate, and adaptive resource management. The conceptual model aims to advance predictive capabilities in reservoir analysis, addressing the complexities of multivariate data integration and opening doors for future innovations in the energy industry. Keywords: Conceptual Model, Multivariate Data, Reservoir Modeling, Petrophysical Analysis.

Registration and estimation for functional generalized linear model

. Functional data analysis has recently concentrated on the challenges of registration and estimation. Our research team has developed an innovative model that enables the simultaneous registration and estimation of multivariate functional data. This approach eliminates the need for pre-smoothing, thereby improving computational efficiency. We utilize the inverse warping function in conjunction with generalized functional principal component analysis (GFPCA) to approximate the covariance structure effectively. The inverse warping function, population mean function, and orthonormal eigenfunctions are estimated using B-spline methods. We have derived the joint probability distribution and established its asymptotic normality. Through extensive simulation studies and real data analyses, we compare our method with existing approaches and demonstrate its exceptional performance, showcasing superior fitting capabilities.

Discovery-Based Analysis for Chemical Trends in Chromatographic Data Sets Using Alteration Analysis and Two-Dimensional Correlation Analysis.

Alteration analysis (ALA), an unsupervised chemometric technique, was evaluated for its ability to discover statistically significant trends in chromatographic data sets. Recently introduced, adoption of ALA has been limited due to uncertainty regarding its sensitivity to minor changes, and there are no rules implementing ALA especially for multivariate data sets such as liquid or gas chromatography coupled to mass spectrometry. Using in-silico data sets, ALA limits of discovery for various signal-to-noises (S/Ns), rates of change across samples, and a number of samples were assessed. For 10 samples, ALA discovered changes of ∼2% across each sample for low S/Ns (15-50), ∼1% change across each sample for moderate S/Ns (65-200), and as little as a 0.1% change at high S/Ns. ALA was also evaluated for unresolved chromatographic peaks, detecting changes down to a resolution of 0.01. In tandem with ALA, two-dimensional correlation analysis (2DCOR), a nonquantitative technique, was employed post-ALA processing to provide unique insights into the relationships between the chemical changes across simulated data sets. Finally, ALA and 2DCOR were applied to the pyrolysis gas chromatography-mass spectrometry (pyGC-MS) of Kraton G1650, a styrene-ethylene-butylene-stryrene (SEBS) polymer, pyrolyzed at temperatures ranging from 350 to 700 °C. A total of 523 statistically significant chemical compounds were discovered. The ALA output was fed into 2DCOR, and a subset of the data was evaluated to understand the relationship between the chemical changes of four selected statistically significant compounds.

Chemometric Methods—A Valuable Tool for Investigating the Interactions Between Antifungal Drugs (Including Antifungal Antibiotics) and Food

Background/Objectives: Developing antifungal drugs with lower potential for interactions with food may help to optimize treatment and reduce the risk of antimicrobial resistance. Chemometrics uses statistical and mathematical methods to analyze multivariate chemical data, enabling the identification of key correlations and simplifying data interpretation. We used the partial least squares (PLS) approach to explore the correlations between various characteristics of oral antifungal drugs (including antifungal antibiotics) and dietary interventions, aiming to identify patterns that could inform the optimization of antifungal therapy. Methods: We analyzed 15 oral antifungal drugs, including azoles (8), antifungal antibiotics (4), antifungal antimetabolites (1), squalene epoxidase inhibitors (1), and glucan synthase inhibitors (1). The input dataset comprised information from published clinical trials, chemical records, and calculations. We constructed PLS models with changes in the pharmacokinetic parameters (∆AUC, area under the curve; ∆Cmax, maximum drug concentration; and ∆Tmax, time to reach maximum drug concentration) after dietary intervention as the response parameters and eight groups of molecular descriptors (M1–M8) as the predictor parameters. We performed separate analyses for the different nutritional interventions. Results: In the final PLS model with food as an intervention, we effectively reduced the dimensionality of the dataset while retaining a substantial percentage of the original information (variance), as significant components explained 69.8% and 17.5% of the predictor and response parameter variances, respectively. The PLS model was significant because its components met the cross-validation criteria. We obtained six significant positive and negative correlations between the descriptors related to atoms and the postprandial ∆Tmax. Conclusions: The PLS method is valuable for investigating interactions between antifungal drugs (including antifungal antibiotics) and food. The correlations obtained can be used in drug modeling to predict interactions with dietary interventions based on the antifungal drug’s chemical structure. Incorporating chemometric techniques into the early drug development stages could facilitate the design of antifungal antibiotics and other antifungal agents with optimized absorption in the presence of dietary components.

Interpretable machine learning-based rolling force prediction using multivariate industrial data during tandem cold rolling

Interpretable machine learning-based rolling force prediction using multivariate industrial data during tandem cold rolling

Otolith chemistry suggests population heterogeneity within a genetically homogeneous Indian scad population along Indian coast

The Indian scad, Decapterus russelli is one of the most exploited pelagic resources of India. Population genetic analyses using mitochondrial and nuclear markers indicated a lack of genetic structuring among populations from Indian waters. As this species is highly migratory, it is also important to establish the environmental influence on its population structure. In the present study, the whole sagittal otolith composition was analysed using Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) to assess spatial distribution of D. russelli collected from 4 sites along the Indian coast. Elemental ratios (Ba/Ca, Fe/Ca, K/Ca, Mg/Ca, Na/Ca, Sr/Ca and Zn/Ca) were analysed using univariate and multivariate statistics to determine whether this chemical signature can provide insight into the adaptive patterns. All element/Ca ratios are found to be significantly different between the four sampling locations (ANOVA, p < 0.05; MANOVA, p < 0.05). Five of the element/Ca ratios were found to be significantly different when the data was analysed coastwise (ANOVA, p < 0.05; MANOVA, p < 0.05). The LDA plot also showed the spatial heterogeneity of Element/Ca ratios between the four sampling sites, but some overlaps were also observed, reflecting the migratory ability of the species. This basic information on spatial ecology is required for formulating effective management and conservation strategies for the species.

Distribution and Source-specific Ecological Risk Analysis of Soil Heavy Metals in South Hunan Province

To identify the spatial distribution patterns and assess the ecological risks associated with soil heavy metal pollution in the southern region of Hunan Province, a total of 362 surface soil samples were collected from the studied area. This study employed multivariate statistics and geographic information systems （GIS） to investigate the spatial distribution pattern of soil metals （Cd, Hg, As, Pb, Zn, Ni, Mn, Tl, and Sb）. Furthermore, the pollution sources and source-specific ecological risk of heavy metals were quantified by combining the positive matrix factorization （PMF） model and the comprehensive ecological risk index model. The results suggested that the mean concentration of soil metals in the study area exceeded the mean background values. Specifically, Cd, Hg, Pb, and Sb showed significant pollution levels and substantial spatial variability, with their concentrations reaching the background values by factors of 8.31, 9.12, 5.09, and 2.79, respectively. The spatial distribution of these heavy metals exhibited discernible patterns, with a higher concentration observed in western Chenzhou and a noticeable trend of increasing concentrations from the central region towards the periphery. The clusters with high concentrations of heavy metals were predominantly situated near industrial zones engaged in black metal smelting and processing and nonferrous metal mining and processing enterprises. Source apportionment indicated that the primary sources of soil heavy metals in the area were industrial activities associated with black metal mining and metallurgy （29.06%）, atmospheric deposition （22.05%）, nonferrous metal mining and processing （20.65%）, a combination of industrial and transportation sources （14.73%）, and natural sources （13.50%）. Moreover, the source-ecological risk assessment model specifically identified Cd and Hg as crucial elements requiring prioritized control measures to mitigate soil ecological risks. Industrial activities, including black metal smelting and processing and atmospheric deposition sources, were identified as the priority sources for health risks in the study region. These results provide an important theoretical basis for the development of effective strategies aimed at managing and diminishing the risks associated with heavy metal contamination in the southern region of Hunan Province.

Spatiotemporal Variation Assessment and Improved Prediction Of Cyanobacteria Blooms in Lakes Using Improved Machine Learning Model Based on Multivariate Data.

Cyanobacterial blooms in shallow lakes pose a significant threat to aquatic ecosystems and public health worldwide, highlighting the urgent need for advanced predictive methodologies. As impounded lakes along the Eastern Route of the South-to-North Water Diversion Project, Lakes Hongze and Luoma play a key role in water resource management, making the prediction of cyanobacterial blooms in these lakes particularly important. To address this, satellite remote sensing data were utilized to analyze the spatiotemporal dynamics of cyanobacterial blooms in these lakes. Subsequently, a precise machine learning model, integrating the Projection Pursuit Model and Random Forest (PP-RF) algorithms, was developed to predict the extent of cyanobacterial blooms, considering a range of influencing factors, including physical, chemical, climatic, and hydrologic variables. The findings indicated pronounced seasonal fluctuations in cyanobacterial blooms, with higher levels in summer than in other seasons. Key determinants for cyanobacterial blooms prediction included solar radiation, temperature and total nitrogen for Lake Hongze, while for Lake Luoma, significant predictors were identified as temperature, water temperature, and solar radiation. Compared with traditional data preprocessing methods, PP-RF model has advantages in addressing multicollinearity. This study provides a feasible method for predicting cyanobacterial blooms in impounded lakes within inter-basin water transfer projects. By inputting region-specific data, this model could be applied broadly, contributing to against the adverse effects of cyanobacterial blooms and provide scientific guidance for the protection and management of aquatic ecosystems.

An adaptive adjustment to the R2 statistic in high-dimensional elliptical models

An adaptive adjustment to the R2 statistic in high-dimensional elliptical models

Differences in the composition of plasma metabolites and intestinal flora of piglets with different weaning weights revealed by untargeted metabolomics and 16S rRNA gene sequencing.

Piglets with different weaning body weights exhibit varying growth performance. This study explores the relationship between their plasma metabolites and gut microbiota to reveal differences in metabolic regulation and microbial composition. Plasma and colon content samples from piglets of different weaning weights were collected. Untargeted metabolomics, 16S rRNA gene sequencing, multivariate statistics, and bioinformatics were used to identify and compare metabolites. Six key findings emerged. First, 23 differential metabolites were found, with three upregulated in high-weight piglets and 20 downregulated in low-weight piglets. A total of 15 were lipids or lipid-like molecules. Second, metabolic pathway enrichment analysis indicated that the sphingolipid signaling pathway, HIF-1 signaling pathway, sphingolipid metabolism pathway, ascorbate and aldarate metabolism pathway, and glycine, serine, and threonine metabolism pathway were the most significantly affected pathways in the plasma of piglets with different weaning body weights. Third, alpha diversity was lower in low-weight weaned piglets. Fourth, Lactobacillus was 23.16% in high-weight piglets, higher than 19.62% in low-weight ones. Fifth, linear discriminant analysis effect size (LEfSe) analysis showed that Faecalibacterium is a biomarker for low-body-weight piglets and Oscillospira is a biomarker for high-body-weight piglets. Finally, Spearman correlation analysis indicated that Lactobacillus, Prevotella, Ruminococcus, and Oscillospira were negatively correlated with differential metabolites in plasma. The plasma metabolites and colon microbiota differed between piglets of different body weights. Lipid-related plasma metabolites contributed to weight variation, being lower in heavier piglets. The colonic microbiota, especially Oscillospira and Roseburia, exhibited strong correlations with these metabolites. © 2025 Society of Chemical Industry.