Data-driven Analysis Research Articles

An improved spectral clustering method for accurate detection of brain resting-state networks

This paper proposes a data-driven analysis method to accurately partition large-scale resting-state functional brain networks from fMRI data. The method is based on a spectral clustering algorithm and combines eigenvector direction selection with Pearson correlation clustering in the spectral space. The method is an improvement on available spectral clustering methods, capable of robustly identifying active brain networks consistent with those from model-driven methods at different noise levels, even at the noise level of real fMRI data.

Data-driven performance analysis of an active chilled beam air conditioning system: A machine learning approach for energy efficiency and predictive maintenance

This study aims to enhance energy performance estimation and maintenance planning for office buildings in Tehran. Machine learning and data mining techniques were employed, utilizing Support Vector Machines (SVMs), Artificial Neural Networks (ANNs), Narrow Neural Networks, and optimizable SVMs. A comprehensive dataset of 2500 experimental samples was utilized with five input parameters (building size, occupancy patterns, environmental conditions, and ACB system parameters) and one output parameter (electricity consumption). GPR models performed better in estimating electricity consumption, achieving a Root Mean Square Error (RMSE) of 297.01 and an impressive R-squared value of 0.99. To optimize energy performance and design more energy-efficient buildings, this model accurately predicts energy consumption. Additionally, this study addressed maintenance planning for ACB systems, simulating maintenance issues with 2500 synthetic samples. The dataset included building age, size, number of floors, usage intensity, previous maintenance history, location, and environmental factors. Maintenance time was estimated using these features, and XGBoost was employed to predict ACB system maintenance time. Low MSE, RMSE, and R-squared values in the XGBoost model allow for accurate predictions. Using this model, facility managers can proactively plan maintenance activities, allocate resources, and minimize downtime. Through proactive measures, ACB performance and maintenance costs are optimized. The study accurately predicts ACB system energy consumption and maintenance time using the rational quadratic GPR model, applicable in Tehran and beyond. Additionally, MATLAB version R2022b was used for prediction, while DesignBuilder and EnergyPlus were utilized for technical analysis.

Recovering biomolecular network dynamics from single-cell omics data requires three time points

Single-cell omics technologies can measure millions of cells for up to thousands of biomolecular features, enabling data-driven studies of complex biological networks. However, these high-throughput experimental techniques often cannot track individual cells over time, thus complicating the understanding of dynamics such as time trajectories of cell states. These “dynamical phenotypes” are key to understanding biological phenomena such as differentiation fates. We show by mathematical analysis that, in spite of high dimensionality and lack of individual cell traces, three time-points of single-cell omics data are theoretically necessary and sufficient to uniquely determine the network interaction matrix and associated dynamics. Moreover, we show through numerical simulations that an interaction matrix can be accurately determined with three or more time-points even in the presence of sampling and measurement noise typical of single-cell omics. Our results can guide the design of single-cell omics time-course experiments, and provide a tool for data-driven phase-space analysis.

Data-driven vulnerability analysis of shared electric vehicle systems to cyberattacks

Disruptions to electric vehicle charging infrastructure can hinder consumers’ confidence and impact the performance of shared electric vehicle (SEV) systems. This study focuses on SEV systems’ vulnerability to emerging cyberattacks at the system level utilizing city-scale SEV trajectory data in Beijing, China. A data-driven simulation model is developed to account for realistic charging demand–supply interactions extracted from the data. The results show that among different types of disruptions, cybersecurity threats present unique challenges to SEV systems. Compared with attacking the most highly utilized charging stations, attacking popular ones on the city’s outskirts is more likely to significantly increase stress on the charging infrastructure and reduce the accessibility of SEVs. Meanwhile, the interactions between SEVs and infrastructure could either amplify the impacts by triggering cascading failures or relieve the adversarial effects. These findings provide insights into designing policies and solutions to enhance the resilience of SEV systems against cyberattacks.

Counterfeit banknote identification based on outliers detection methods

Counterfeit banknotes in circulation in the economy are a long-standing and recurring problem. Technological improvements are applied to produce increasingly convincing counterfeit banknotes. In contrast, sophisticated strategies to identify counterfeit notes remain an up-to-date and fascinating research topic of great interest. The literature presents several techniques for detecting counterfeit banknotes. In this study, outlier detection methodologies are presented as adapted mechanisms to assist in the fight against counterfeit banknotes. This study addresses and tests multivariate outlier detection techniques' efficiency in identifying counterfeit banknotes. A comprehensive computational experiment proves the great adaptability of the methods to multivariate outliers in the identification of counterfeit banknotes. In particular, the Data-driven Cluster Analysis Method (DDCAM) technique presented promising and optimistic results through several quality and performance measurement metrics, suggesting a hopeful future for detecting counterfeit banknotes.

Data-driven analysis of heating and cooling base temperatures for buildings: Case studies in South Korea

To investigate statistical differences in the heating (Tbh) and cooling base temperatures (Tbc) across various building types, vintages, and sizes, we conducted case studies in Seoul, South Korea, based on data from more than 10,000 buildings. To assess the relative impacts of internal and external loads on Tbh and Tbc, we used a core-to-perimeter zone ratio approximated by the average floor area, defined as the floor area per building floor. Employing the energy signature method, change-point regression, and calendarization, we determined the best-fit change-point model and the corresponding Tbh and Tbc for each building. Our findings revealed statistically significant differences in Tbhs and Tbcs, with variations of up to 6 and 4 °C, respectively, across six building types. Notably, for most building types, Tbh and Tbc decreased as the average floor area increased, possibly influenced by a higher core zone to total area ratio. This highlights the predominant impact of internal loads over external loads. The influence of the average floor area on Tbh and Tbc outweighed that of building vintage, particularly for lodging, medical, and sales facilities, where Tbh and Tbc decreased in the more recent vintage buildings. However, this influence was less pronounced for facilities for senior citizens and children and education and research facilities, possibly due to factors such as diverse building shapes and the thermal preferences of occupants. Although this study focuses on South Korea, our data-driven approach provides statistical evidence for refining base temperature and improving traditional degree-day methods. Our findings suggest the need to establish base temperatures that consider various building types, vintages, and average floor areas, especially for large metropolitan cities where these factors vary widely.

Unsupervised Learning of Eye State Prototypes for Semantically Rich Blinking Detection.

Blinking contributes to the health and protection of the eye and also holds potential in the context of muscle or nerve disorder diagnosis. Traditional methods of classifying eye blinking as open or closed are insufficient, as they do not capture medical-relevant aspects like closure speed, duration, or percentage. The issue could be solved by reliably detecting blinking intervals in high-temporal recordings. Our research demonstrates the reliable detection of blinking events through data-driven analysis of the eye aspect ratio. In an unsupervised manner, we establish an eye state prototype to identify blink intervals and measure inter-eye synchronicity between moments of peak closure. Additionally, our research shows that manually defined prototypes yield comparable results. Our results demonstrate inter-eye synchronicity up to 4.16 ms. We anticipate that medical professionals could utilize our methods to identify or define disease-specific prototypes as potential diagnostic tools.

Pre-service teachers co-constructing narratives about the future of education

ABSTRACT This study examined the kinds of narratives that nine pre-service teachers shared in online collaborative learning discussions about the future of education in relation to global megatrends, namely digitalisation and ecological sustainability. We also analysed how they positioned themselves in the future of education. We used data-driven qualitative analysis and narrative analysis, and we report our findings partly as non-fiction comic strips. We found that the pre-service teachers (1) viewed digitalisation in education through antonyms and ambivalence, (2) emphasised critical media literacy, (3) viewed ecological perspectives through a main and counter-narrative (taking root more deeply vs. not everyone needs to get excited) and (4) emphasised the role of action. In terms of positioning, we found dynamic tensions between passive and active stances. We discuss our findings in light of teacher identity and education for democracy and sustainability. The results can be used as thinking tools in teacher education.

Cerebro-spinal somatotopic organization uncovered through functional connectivity mapping

Abstract Somatotopy, the topographical arrangement of sensorimotor pathways corresponding to distinct body parts, is a fundamental feature of the human central nervous system (CNS). Traditionally, investigations into brain and spinal cord somatotopy have been conducted independently, primarily utilizing body stimulations or movements. To date, however, no study has probed the somatotopic arrangement of cerebro-spinal functional connections in vivo in humans. In this study, we used simultaneous brain and cervical spinal cord functional magnetic resonance imaging (fMRI) to demonstrate how the coordinated activities of these two CNS levels at rest can reveal their shared somatotopy. Using functional connectivity analyses, we mapped preferential correlation patterns between each spinal cord segment and distinct brain regions, revealing a somatotopic gradient within the cortical sensorimotor network. We then validated this large-scale somatotopic organization through a complementary data-driven analysis, where we effectively identified spinal cord segments through the connectivity profiles of their voxels with the sensorimotor cortex. These findings underscore the potential of resting-state cerebro-spinal cord fMRI to probe the large-scale organization of the human sensorimotor system with minimal experimental burden, holding promise for gaining a more comprehensive understanding of normal and impaired somatosensory-motor functions.

Machine-learning assisted analysis on the seismic performance of steel reinforced concrete composite columns

This paper presents the data-driven analysis on the seismic performance of steel reinforced concrete (SRC) composite columns by using machine learning (ML) algorithms. A total of 248 test data of SRC columns subjected to the combined axial compressive force and low reversed cyclic horizontal force was collected from the published literatures to form the database. Since the seismic action triggers the SRC columns to suffer from different failure mode and hence exhibit quite different load bearing mechanism, six ML algorithms were employed to facilitate the failure mode classification and bearing capacity prediction. It was found that the Random Forest (RF) model predicts the failure mode most accurately, while XGBoost model delivers the best estimation of bearing capacity. To further evaluate the feasibility of ML models and unveil the interaction between different variables, interpretability analysis of ML models was carried out. In comparison with the conventional empirical judgement and theoretical derivation, ML model delivers enhanced accuracy and robustness to predict the seismic performance of SRC columns, and hence it could be used as a promising alternative for the preliminary design and analysis of SRC columns.

Brain transcriptomic signatures for mood disorders and suicide phenotypes: an anterior insula and subgenual ACC network postmortem study.

Mood disorders affect over ten percent of humans, but studies dissecting the brain anatomical and molecular neurobiological mechanisms underlying mood (dys)functions have not consistently identified the patterns of pathological changes in relevant brain regions. Recent studies have identified pathological changes in the anterior insula (Ant-Ins) and subgenual anterior cingulate (sgACC) brain network in mood disorders, in line with this network's role in regulating mood/affective feeling states. Here, we applied whole-tissue RNA-sequencing measures of differentially expressed genes (DEGs) in mood disorders versus (vs.) psychiatrically unaffected controls (controls) to identify postmortem molecular pathological markers for mood disorder phenotypes. Using data-driven factor analysis of the postmortem phenotypic variables to determine relevant sources of population variances, we identified DEGs associated with mood disorder-related diagnostic phenotypes by combining gene co-expression, differential gene expression, and pathway-enrichment analyses. We found downregulation/under expression of inflammatory, and protein synthesis-related genes associated with psychiatric morbidity (i.e., all co-occurring mental disorders and suicide outcomes/death by suicide) in Ant-Ins, in contrasts to upregulation of synaptic membrane and ion channel-related genes with increased psychiatric morbidity in sgACC. Our results identified a preponderance of downregulated metabolic, protein synthesis, inflammatory, and synaptic membrane DEGs associated with suicide outcomes in relation to a factor representing longevity in the Ant-Ins and sgACC (AIAC) network. Our study revealed a critical brain network molecular repertoire for mood disorder phenotypes, including suicide outcomes and longevity, and provides a framework for defining dosage-sensitive (i.e., downregulated vs. upregulated) molecular signatures for mood disorder phenotypic complexity and pathological outcomes.

Data-driven risk analysis of nonlinear factor interactions in road safety using Bayesian networks

This paper aims to demonstrate nonlinear risk factor interactions based on a data-driven approach using a Bayesian network model, providing a road safety use case. Road safety is a critical issue worldwide, with approximately 1.3 million road traffic deaths each year (WHO). Traditional road safety risk assessment methods often analyze individual factors separately; however, these assessments fail to capture the complex dynamics of real-world analysis, in which multiple factors interact through nonlinear relationships. In this study, a novel road safety risk assessment approach that uses a Bayesian network model to explore the nonlinear relationships among road safety risk factors is developed. Through the analysis of extensive crash reports from the state of Maryland, the complex interdependencies among various risk factors and their cumulative impact on road safety are investigated. Our findings show that two combined risk factors have different effects on risk level when considered individually. Two case studies related to human state risk factors and environmental risk factors, such as driving under the influence and snowy roads, as well as fatigue and snowy roads, have an amplified effect on the risk level. The findings highlight the importance of considering nonlinear interactions among risk factors when developing effective and targeted strategies for accident prevention and road safety improvement. This research contributes to the field of road safety by presenting a new methodology for understanding and mitigating road safety hazards.

Factors influencing the tendency of residual symptoms in patients with depressive disorders: a longitudinal study

BackgroundResidual symptoms of depressive disorders are serious health problems. However, the progression process is hardly predictable due to high heterogeneity of the disease. This study aims to: (1) classify the patterns of changes in residual symptoms based on homogeneous data, and (2) identify potential predictors for these patterns.MethodsIn this study, we conducted a data-driven Latent Class Growth Analysis (LCGA) to identify distinct tendencies of changes in residual symptoms, which were longitudinally quantified using the QIDS-SR16 at baseline and 1/3/6 months post-baseline for depressed patients. The association between baseline characteristics (e.g. clinical features and cognitive functions) and different progression tendencies were also identified.ResultsThe tendency of changes in residual symptoms was categorized into four classes: “light residual symptom decline (15.4%)”, “residual symptom disappears (39.3%)”, “steady residual symptom (6.3%)” and “severe residual symptom decline (39.0%)”. We observed that the second class displayed more favorable recuperation outcomes than the rest of patients. The severity, recurrence, polypharmacy, and medication adherence of symptoms are intricately linked to the duration of residual symptoms’ persistence. Additionally, clinical characteristics including sleep disturbances, depressive moods, alterations in appetite or weight, and difficulties with concentration have been identified as significant factors in the recovery process.ConclusionsOur research findings indicate that certain clinical characteristics in patients with depressive disorders are associated with poor recovery from residual symptoms following acute treatment. This revelation holds significant value in the targeted attention to specific patients and the development of early intervention strategies for residual symptoms accordingly.

(Invited) Battery Manufacturing Needs from the Perspective of DOE's Advanced Materials and Manufacturing Technologies Office

The Department of Energy’s Advanced Materials and Manufacturing Technologies Office (AMMTO) continues to invest in energy storage research, development, and demonstration (RD&D) to help stakeholders improve efficiency, cut costs, and make materials, components, and systems with superior performance. Despite recent promising advances, manufacturing capabilities are still necessary to meet the expected demand for batteries as we move toward a clean energy economy. Strengthening the domestic manufacturing supply chains is also another important task to pursue in parallel. A multi-pronged approach is required to strengthen domestic battery manufacturing and reduce time to commercialization and deployment of transformative technologies. AMMTO recognizes a number of unmet needs that need to be addressed including development of manufacturing processes and equipment that take advantage of the most recent technical advances as well as effective utilization of data-driven analysis and experimentation to assist in manufacturing optimization.In this talk, AMMTO’s perspective on energy storage technology development and manufacturing will be discussed, highlighting current gaps and challenges that need to be addressed. In addition, there will be a robust discussion of a wide variety of DOE’s efforts in the context of technical and manufacturing challenges regarding scale-up and performance that still prevent the electrochemical energy storage community from achieving cost targets and commercial viability.

Functional network characteristics in anxiety- and mania-based subgroups of bipolar I disorder

BackgroundBipolar disorder I (BD-I) is a heterogeneous disorder with a high prevalence of comorbid anxiety. The aim of this study was to investigate whether anxiety and mania symptoms define distinct subgroups within BD-I and to explore potential differences in functional network characteristics between these subgroups. MethodsSubgroups were identified using scores from clinical anxiety and mania scales. After dimension reduction of these scores, data-driven clustering analysis with cross-validation was employed to reveal the existence of subgroups. Resting-state functional magnetic resonance imaging (rs-fMRI) scans were pre-processed using fMRIPrep. After parcellation and network construction, global and regional graph theoretical measures were calculated per subgroup. ResultsClustering results revealed that, based on anxiety symptomatology, subjects fell into two distinct subgroups, whereas mania symptoms divided subjects into four unique subgroups. These subgroups varied notably on several symptom scales. Network assortativity was significantly associated with anxiety subgroups. Post-hoc pairwise comparisons did not reveal significant global functional network differences between the anxiety subgroups or between mania subgroups. Regional network differences between clinical subgroups were especially apparent for strength and degree in the temporal and frontal lobes. LimitationsSmall sample size of some subgroups is a limitation of this study as is the categorical rather than continuous representation of anxiety and mania symptoms. ConclusionsBD-I populations may be stratified into robust subgroups based on anxiety and mania symptoms, showing differences in functional network connectivity. Our findings highlight new avenues of research for investigating heterogeneity in psychiatric populations.

Characteristic value decomposition: A unifying paradigm for data-driven modal analysis

Data-driven modal analysis is an indispensable means to understand the dynamic behavior of engineered structures and natural systems. It effectively captures active dynamic behavior and embeds them in a set of identified linear modal subspaces, unveiling complex dynamic behavior via the superposition of simpler hierarchical spatiotemporal dynamics. Despite the existence of many methods, there is no unified perspective connecting these modal analysis techniques. This paper proposes a new modal decomposition method, the characteristic value decomposition (CVD), that provides a unifying paradigm for data-driven modal analysis. CVD identifies identical spatial modes and temporal coordinates from two measurement data matrices. The modal parameters are identified through an eigendecomposition to a quotient of auto-covariance and cross-covariance matrices of the two data matrices. This paper shows that several data-driven modal analysis methods, namely, the state-variable decomposition, the dynamic mode decomposition, the Ibrahim time domain method, and the eigensystem realization algorithm, are special cases of CVD when the two data matrices are appropriately selected. Moreover, the CVD identifies the modal parameters with higher accuracy and is robust to added noise compared to established methods, as demonstrated using numerical experiments. It is also shown that the CVD flexibly utilizes data from multiple experiments and undersampled data sets to unveil system dynamics.

Correlation Among Neighborhood-level Measures of the Tobacco Retail Environment.

Tobacco retailer density and distance to tobacco retailers are understood to influence tobacco-related behaviors; however, there is no general agreement on how to best characterize or measure the tobacco retail environment (TRE). In this data-driven analysis, we examine similarities among neighborhood-level measures of the TRE and assess how geographic resolution of the neighborhood units may affect them. We used locations of likely tobacco retailers in the United States (US) to calculate multiple retailer count, density, and distance measures. Measures were calculated at the Census block group, tract, and county (including county equivalents) levels of geographic aggregation. Spearman's correlation was used to evaluate similarity among the TRE measures. At the block group and tract level, correlation among all TRE measures ranged from slightly negative (ρ = -0.03) to nearly perfect (ρ = 0.99). At both levels of aggregation, distance-based TRE measures were highly correlated (ρ > 0.76). At the block group level, simple count of retailers was highly correlated with the density measures (ρ > 0.83), and at the tract level, simple count was moderately to highly correlated with the density measures (ρ > 0.5). Findings were generally similar at the county level; a notable deviation was that retailers per person was negatively correlated with all other TRE measures (range from ρ = -0.08 to ρ = -0.32). Some common measures were not correlated, suggesting they capture different aspects of the TRE; similarity among the various measures also varied by level of geographic aggregation. Because the tobacco retail environment shapes people's tobacco-related behaviors, using appropriate measures to characterize it at a neighborhood-level is paramount. Our work highlights both the similarities and differences among a set of common measures, thereby suggesting the measures may be capturing different aspects of the overall retail environment. Our findings regarding geographic level of aggregation underscore the importance of neighborhood definition in any tobacco retail environment analysis.

Focus on the disruption of networks and system dynamics.

Networks are designed to ensure proper functioning and sustained operability of the underlying systems. However, disruptions are generally unavoidable. Internal interactions and external environmental effects can lead to the removal of nodes or edges, resulting in unexpected collective behavior. For instance, a single failing node or removed edge may trigger a cascading failure in an electric power grid. This Focus Issue delves into recent advances in understanding the impacts of disruptions on networks and their system dynamics. The central theme is the disruption of networks and their dynamics from the perspectives of both data-driven analysis as well as modeling. Topics covered include disruptions in the dynamics of empirical systems such as nuclear reaction networks, infrastructure networks, social networks, epidemics, brain dynamics, and physiology. Emphasis is placed on various phenomena in collective behavior, including critical phase transitions, irregular collective dynamics, complex patterns of synchrony and asynchrony, chimera states, and anomalous oscillations. The tools used for these studies include control theory, diffusion processes, stochastic processes, and network theory. This collection offers an exciting addition to the evolving landscape of network disruption research.

Estimation of Carleman operator from a univariate time series.

Reconstructing a nonlinear dynamical system from empirical time series is a fundamental task in data-driven analysis. One of the main challenges is the existence of hidden variables; we only have records for some variables, and those for hidden variables are unavailable. In this work, the techniques for Carleman linearization, phase-space embedding, and dynamic mode decomposition are integrated to rebuild an optimal dynamical system from time series for one specific variable. Using the Takens theorem, the embedding dimension is determined, which is adopted as the dynamical system's dimension. The Carleman linearization is then used to transform this finite nonlinear system into an infinite linear system, which is further truncated into a finite linear system using the dynamic mode decomposition technique. We illustrate the performance of this integrated technique using data generated by the well-known Lorenz model, the Duffing oscillator, and empirical records of electrocardiogram, electroencephalogram, and measles outbreaks. The results show that this solution accurately estimates the operators of the nonlinear dynamical systems. This work provides a new data-driven method to estimate the Carleman operator of nonlinear dynamical systems.

ModsNet: Performance-Aware Top- k Model Search Using Exemplar Datasets

We demonstrate ModsNet , a search tool for pre-trained data science MOD el s recommendatio N using E xamplar da T aset. Given a set of pre-trained data science models, an "example" input dataset, and a user-specified performance metric, ModsNet answers the following query: "what are top-k models that have the best expected performance for the input data?" The need for searching high-quality pre-trained models is evident in data-driven analysis. Inspired by "query by example" paradigm, ModsNet does not require users to write complex queries, but only provide an "examplar" dataset, a task description, and a performance measure as input, and can automatically suggest top- k matching models that are expected to have desirable performance to perform the task over the provided sample dataset. ModsNet utilizes a knowledge graph to integrate model performances over datasets and synchronizes it with a bipartite graph neural network to estimate model performance, reduce inference cost, and promptly respond to top- k model search queries. To cope with strict cold-start (upon receiving a new dataset when no historical performance of registered models are observed), it performs a dynamic, cost-bounded "probe-and-select" strategy to incrementally identify promising models. We demonstrate the application of ModsNet in enabling efficient scientific data analysis.