Time-varying Data Research Articles

3D Based Visualization Tool to Analyze the Influential Topics via Hashtags on Instagram Platform

This paper intends to develop an interactive, comprehensive information visualization platform of Instagram hashtag analysis. Instagram hashtags has developed themselves into all different kinds of group or communities for users to share hobbies and find similar friends. In order to analyze topic influence and user interest trend from Instagram, which contains billions of end-users and has worldwide influence, hashtag analysis is necessary to gather such information and compare the proportion of people involving in each tags and rank them to visualize. The visualization is developed in 3D space and consists of time-varying data flow of tags, together with tag comparison analysis, as well as event researches. In the rest of the paper, we mainly discuss the design idea and the development process of the system. An example of the system design work will be shown in the discussion, which involves 4 popular hashtags discussed on Instagram and are shown on the system, displayed as an 3D histogram, together with another comparison histogram to compare different tags, as well as an event view in the back.

Unilateral secession, international recognition, and great power contestation

Recognition of aspiring states from established countries is central to becoming a member state of the international system. Previous research suggests that great power recognition decisions regarding aspiring states rapidly converge toward either recognition or non-recognition, yet great power convergence has still not occurred in the case of Kosovo after more than ten years. Unilateral secessions typically remain wholly unrecognized, since they violate the norm of home state consent, yet Kosovo has now been recognized by more than 100 countries. Why do some countries extend recognition to unilateral secessions, and do so early, whereas others delay recognition or withhold it altogether? In the case of Kosovo, great power influence and contestation, rather than convergence, have played a key role in shaping recognition decisions. We argue that countries in the US sphere of influence, with strong economic and military ties, are more likely to recognize Kosovo and to do so relatively fast, whereas countries influenced by Russia are less likely to recognize Kosovo at all, or to do so only after an extended delay. However, great powers are not equal in influencing other states to adopt their preferred position, since the USA is more powerful than Russia and can benefit from working alongside allies within the Western-oriented world order. We estimate a non-proportional Cox model with new time-varying data on Kosovo recognition and provide evidence that US military ties influenced other countries in extending recognition to unilateral secession.

PCV93 Using Machine Learning to Predict Anticoagulation Control in Atrial Fibrillation: A UK Retrospective Database Study

To investigate the predictive performance of machine learning (ML) algorithms for estimating anticoagulation control (AC) in patients with atrial fibrillation (AF), treated with vitamin K antagonists. This was a retrospective cohort study of adult patients (≥18 years) between 2007 and 2016 using linked primary and secondary care data (Clinical Practice Research Datalink GOLD and Hospital Episode Statistics). Various ML techniques were explored to predict poor AC, defined as time in therapeutic range (TTR) <70% based on international normalized ratio (INR) 2.0-3.0. Baseline (linear and non-linear support vector machines; random forests; stochastic gradient boosting [XGBoost]; neural networks [NN]) and time-varying data (6 week intervals up to 30 weeks (long-short term memory [LSTM] NN)) were employed. Patient records representing unique lines of warfarin therapy (LOT) were separated into training (70%) and holdout sets (30%) for model training and testing. 35,479 patients were eligible for inclusion, of whom 24,684 and 10,795 were assigned to the training (32,683 unique LOTs) and holdout sets (14,218 unique LOTs). Across all models, depression was a significant driver in predicting AC. At baseline, XGBoost was the best-performing model (area under the curve [AUC]: 0.624) due to its ability to identify non-linear interactions for factors including age and weight (greater probability of poor control: <65 and >80 years and <70kg, respectively). Addition of time-varying data to the LSTM NN improved predictive performance, plateauing at AUC of 0.830 at 30 weeks. ML algorithms displayed very good ability in predicting patients at greater risk of poor AC. The addition of time-varying data to the algorithm, especially prior INR measurements, improved predictive performance. These algorithms may be clinically valuable in identifying and supporting patients who may benefit from more frequent INR monitoring or switching to therapies not requiring dose adjustments.

Classification of drug use patterns.

Characterizing long‐term prescription data is challenging due to the time‐varying nature of drug use. Conventional approaches summarize time‐varying data into categorical variables based on simple measures, such as cumulative dose, while ignoring patterns of use. The loss of information can lead to misclassification and biased estimates of the exposure‐outcome association. We introduce a classification method to characterize longitudinal prescription data with an unsupervised machine learning algorithm. We used administrative databases covering virtually all 1.3 million residents of Manitoba and explicitly designed features to describe the average dose, proportion of days covered (PDC), dose change, and dose variability, and clustered the resulting feature space using K‐means clustering. We applied this method to metformin use in diabetes patients. We identified 27,786 metformin users and showed that the feature distributions of their metformin use are stable for varying the lengths of follow‐up and that these distributions have clear interpretations. We found six distinct metformin user groups: patients with intermittent use, decreasing dose, increasing dose, high dose, and two medium dose groups (one with stable dose and one with highly variable use). Patients in the varying and decreasing dose groups had a higher chance of progression of diabetes than other patients. The method presented in this paper allows for characterization of drug use into distinct and clinically relevant groups in a way that cannot be obtained from merely classifying use by quantiles of overall use.

Test for the covariance matrix in time-varying coefficients panel data models with fixed effects

This paper proposes tests for the null of sphericity and identity matrix for nonparametric time-varying coefficient panel data models with fixed effects. Firstly, based on the local linear smoothing technique, the estimators of the unknown coefficient functions and model residuals are obtained. Secondly, proper test statistics are proposed aiming at tests for sphericity or identity matrix with a large number of cross-sectional units and time series observations. In addition, the limiting distributions of the proposed test statistics are derived based on random matrix theory. At last, some simulation studies are conducted to examine the finite sample performance for the proposed test statistics and a real data example is analyzed.

Outdoor light at night and risk of coronary heart disease among older adults: a prospective cohort study.

We estimated the association between outdoor light at night at the residence and risk of coronary heart disease (CHD) within a prospective cohort of older adults in Hong Kong. Over a median of 11 years of follow-up, we identified 3772 incident CHD hospitalizations and 1695 CHD deaths. Annual levels of outdoor light at night at participants' residential addresses were estimated using time-varying satellite data for a composite of persistent night-time illumination at ∼1 km2 scale. We used Cox proportional hazards models to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) of the association between outdoor light at night at the residence and risk of CHD. The association between light at night and incident CHD hospitalization and mortality exhibited a monotonic exposure-response function. An interquartile range (IQR) (60.0 nW/cm2/sr) increase in outdoor light at night was associated with an HR of 1.11 (95% CI: 1.03, 1.18) for CHD hospitalizations and 1.10 (95% CI: 1.00, 1.22) for CHD deaths after adjusting for both individual and area-level risk factors. The association did not vary across strata of hypothesized risk factors. Among older adults, outdoor light at night at the residence was associated with a higher risk of CHD hospitalizations and deaths. We caution against causal interpretation of these novel findings. Future studies with more detailed information on exposure, individual adaptive behaviours, and potential mediators are warranted to further examine the relationship between light at night and CHD risk.

AI for Beyond 5G Networks: A Cyber-Security Defense or Offense Enabler?

Artificial intelligence (AI) is envisioned to play a pivotal role in empowering intelligent, adaptive and autonomous security management in 5G and beyond networks, thanks to its potential to uncover hidden patterns from a large set of time-varying multi-dimensional data, and deliver faster and accurate decisions. Unfortunately, AI's capabilities and vulnerabilities make it a double-edged sword that may jeopardize the security of future networks. This article sheds light on how<;span class="upper-case"> Ai <;/span>may impact the security of 5G and its successor from its posture of defender, offender or victim, and recommends potential defenses to safeguard from malevolent AI while pointing out their limitations and adoption challenges.

Association Between Outdoor Light at Night and Risk of Coronary Heart Disease among 60,229 Older Adults: A Prospective Cohort Study

Background Light at night (LAN) is the most important environmental cue for circadian misalignment, which may increase risk of coronary heart disease through impairing circadian rhythms. We aimed to estimate the association between exposure to outdoor LAN and coronary heart disease (CHD) in elders.MethodsWe followed 60,229 Hong Kong elders free of CHD from 1998 through 2011. Annual LAN exposure was estimated using time-varying satellite data for a composite of persistent nighttime illumination at ~1 km2 scale for each residence. We used Cox proportional hazard models to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between exposure to outdoor LAN and risk of CHD. Results Over 10 to 13 years of follow-up, we identified 3,772 incident CHD cases and 1,695 CHD deaths. Annual levels of LAN at participants’ residential addresses ranged from 2.2 to 233.1 nW/cm2/sr. The concentration-response functions for the association between LAN and CHD incidence and mortality were approximate linear with the slopes becoming flat when LAN above 110 nW/cm2/sr. Per interquartile range (IQR) (60.0 nW/cm2/sr) increase in outdoor LAN, the HR was 1.09 (95% CI: 1.03, 1.17) for CHD incidence and 1.10 (95% CI: 1.01, 1.22) for CHD mortality after adjusting for both individual and area-level risk factors.Conclusions Exposure to outdoor light at night was associated with higher risk of coronary heart disease among Hong Kong elders. Our findings highlight the need to regulate light pollution for the potential coronary heart disease benefits.

Clustering for time-varying relational count data

Relational count data are often obtained from sources such as simultaneous purchase in online shops and social networking service information. Clustering such relational count data reveals the latent structure of the relationship between objects such as household items or people. When relational count data observed at multiple time points are available, it is worthwhile incorporating the time structure into the clustering result to understand how objects move between the clusters over time. In this paper, we propose two clustering methods for analyzing time-varying relational count data. The first model, the dynamic Poisson infinite relational model (dPIRM), handles time-varying relational count data. In the second model, which we call the dynamic zero-inflated Poisson infinite relational model, we further extend the dPIRM so that it can handle zero-inflated data. Proposing both two models is important as zero-inflated data are often encountered, especially when the time intervals are short. In addition, by explicitly deriving the relevant full conditional distributions, we describe the features of the estimated parameters and, in turn, the relationship between the two models. We show the effectiveness of both models through a simulation study and a real data example.

Condition-Based Monitoring in Variable Machine Running Conditions Using Low-Level Knowledge Transfer With DNN

Traditional machine learning methods assume that training and testing data must be from the same machine running condition (MRC) and drawn from the same distribution. However, in several real-time industrial applications, this assumption does not hold. The traditional methods work satisfactorily in steady-state conditions but fail in time-varying conditions. In order to utilize time-varying data in variable MRCs, this article proposes a novel low-level knowledge transfer framework using a deep neural network (DNN) model for condition monitoring of machines in variable running conditions. The low-level features have been extracted in time, frequency, and time–frequency domains. These features are extracted from the source data to train the DNN. The trained DNN-based parameters are then transferred to another DNN, which is modified according to the low-level features extracted from the target data. The proposed approach is validated through three case studies on: 1) the air compressor acoustic data set; 2) the Case Western Reserve University bearing data set; and 3) the intelligent maintenance system bearing data set. The prediction accuracy obtained for the above case studies is as high as 100%, 93.07%, and 100%, respectively, with fivefold cross-validation. These real-time results show considerable improvement in the prediction performance using the proposed approach. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —Condition-based monitoring schemes are widely applicable to rotating machines in various industries since they operate in tough working situations, and consequently, unpredicted failures occur. These unpredicted failures may cause perilous accidents in the industries. CBM systems prevent such failures, which results in the reduction of equipment damage and, hence, increases machinery lifetime. Modern industries are so complex and generating huge data, and these data can be collected using sensors, but placing a large number of sensors is difficult and expensive for different but similar kinds of faults in industries. This also increases the cost due to additional sensors and circuits. In this article, the authors have proposed a novel low-level knowledge transfer framework using the deep neural network (DNN)-based method for condition monitoring of machines in variable running conditions. Low-level features have been extracted to reduce the computations of DNN drastically with improved performance. This article also considered additional faults in the target domain, which is more practical in real-time applications. The proposed scheme has been validated with three case studies on acoustic and vibration signatures.

Web-based 3D visualization framework for time-varying and large-volume oceanic forecasting data using open-source technologies

There are limitations to traditional visualization solutions regarding real-time 3D visualization of time-varying and large-volume 3D gridded oceanographic data in a web environment. We adopted the open-source visualization technologies to implement a browser-based 3D visualization framework. The developed 3D visualization interfaces provide users 3DGIS experiences on a virtual globe and simultaneously provide efficient 3D volume rendering and enriched interactive volume analysis. Our experiments suggest that the well-designed Cesium and Plotly.js API allow researchers to easily establish 3D visualization applications while avoiding the requirements of intensive programming and computations. The case study conducted shows that the proposed methods is a feasible alternative web-based 3D visualization solution, which provides a faster rendering speed, high visual effects and on-the-fly 3D visualization of oceanographic data. Due to its open-source architecture and the simplicity of the adopted technologies, the visualization framework can be easily customized to visualize other scientific data with few modifications.

Does Economic Growth Driving Force Convert? Evidence from China

The conversion from conventional to new driving forces and their time-varying characteristics are of great importance to China’s economic transformation. In this paper, we attempt to investigate the economic growth driving force conversion in China and the time-varying characteristics of driving forces by constructing a time-varying coefficient panel data model during the period of 1998–2015. The empirical results indicate that China’s economy has undergone driving force conversion. Specifically, China’s economic growth driving forces have been transformed from traditional ones (human capital and gross fixed capital formation) to new ones (innovation capacity and structural transformation). Furthermore, we find that the characteristics of the driving forces are time-varying and heterogeneous. Innovation capacity and structural transformation have a more crucial impact on economic growth. Finally, based on the conclusions of the quantitative analysis, some important policy implications can be pursued to foster economic growth. Chinese government ought to enact various policies that are conducive to enhancing innovation capacity and accelerating structural transformation.

Long-term determinants of income inequality: evidence from panel data over 1870–2016

We investigate the major factors which drive income inequality in the OECD countries, using long panel data which span the period from 1870 to 2016. We consider two measures of inequality: the Gini coefficient and the income share of the top 10% of the population. Employing the panel vector auto-regression method, we show that a positive shock to the real interest rate and government spending is negatively and significantly associated with income inequality in the middle, as well as at the top end of the income distribution. An increase in real GDP per capita leads to an increase in income inequality, measured by the Gini coefficient, whereas an advance in financial development reduces it. We find that income inequality responds negatively to positive innovation shocks initially, but this effect becomes positive with some time lag for top-income inequality. Educational attainment significantly reduces top-income inequality. Our results are robust to alternative specifications, including the local projection method and estimations based on different samples. We also capture the time dynamics in our series using a time-varying nonparametric panel data model and show that the real interest rate and financial development are negatively associated with income inequality for most of the period in the post-World War II era, while the effect of real GDP per capita is positive and significant over the same period.

DDxNet: a deep learning model for automatic interpretation of electronic health records, electrocardiograms and electroencephalograms

Effective patient care mandates rapid, yet accurate, diagnosis. With the abundance of non-invasive diagnostic measurements and electronic health records (EHR), manual interpretation for differential diagnosis has become time-consuming and challenging. This has led to wide-spread adoption of AI-powered tools, in pursuit of improving accuracy and efficiency of this process. While the unique challenges presented by each modality and clinical task demand customized tools, the cumbersome process of making problem-specific choices has triggered the critical need for a generic solution to enable rapid development of models in practice. In this spirit, we develop DDxNet, a deep architecture for time-varying clinical data, which we demonstrate to be well-suited for diagnostic tasks involving different modalities (ECG/EEG/EHR), required level of characterization (abnormality detection/phenotyping) and data fidelity (single-lead ECG/22-channel EEG). Using multiple benchmark problems, we show that DDxNet produces high-fidelity predictive models, and sometimes even provides significant performance gains over problem-specific solutions.

A Temporal Multidimensional Model and OLAP Operators

Usually, data in data warehouses (DWs) are stored using the notion of the multidimensional (MD) model. Often, DWs change in content and structure due to several reasons, like, for instance, changes in a business scenario or technology. For accurate decision-making, a DW model must allow storing and analyzing time-varying data. This paper addresses the problem of keeping track of the history of the data in a DW. For this, first, a formalization of the traditional MD model is proposed and then extended as a generalized temporal MD model. The model comes equipped with a collection of typical online analytical processing (OLAP) operations with temporal semantics, which is formalized for the four classic operations, namely roll-up, dice, project, and drill-across. Finally, the mapping from the generalized temporal model into a relational schema is presented together with an implementation of the temporal OLAP operations in standard SQL.

Temporal flow of hubs and connectivity in the human brain

Hubs in brain network connectivity have previously been observed using neuroimaging techniques and are generally believed to be of pivotal importance to establish and maintain a functional platform on which cognitively meaningful and energy-efficient neuronal communication can occur. However, little is known if hubs are static (i.e. a brain region is always a hub) or if these properties change over time (i.e. brain regions fluctuate in their ‘hubness’). To address this question, we introduce two new methodological concepts, the flow of brain connectivity and node penalized shortest paths which are then applied to time-varying functional connectivity fMRI BOLD data. We show that the constellations of active hubs change over time in a non-trivial way and that activity of hubs is dependent on the temporal scale of investigation. Slower fluctuations in the number of active hubs that exceeded the degree expected by chance alone were detected primarily in subcortical structures. Moreover, we observed faster fluctuations in hub activity residing predominately in the default mode network that suggests dynamic events in brain connectivity. Our results suggest that the temporal behavior of connectivity hubs is a multilayered and complex issue where method-specific properties of temporal sensitivity to time-varying connectivity must be taken into account. We discuss our results in relation to the on-going discussion of the existence of discrete and stable states in the resting-brain and the role of network hubs in providing a scaffold for neuronal communication across time.

Visualization and Analysis of the Volume Dataset for Time‐Varing Electromagnetic Simulation

The three Dimensional (3D) visualization of time-varying Electromagnetic (EM) data in antenna devices can be used to evaluate various parameters, such as the quality of beams and the power performance of antennas, for optimizing the antenna design. Compared with the single volume data, the time-varying data has characteristics of strong continuity, large scale and multi-variable. It is challenging. We proposed a real-time dynamic rendering method using Compute unified device architecture (CUDA)-based volume rendering algorithm. We designed the opacity transfer function as multiple trapezoid, which represents each energy band, and provided a fusion strategy for a horn-like surface and a EM volume model that considers the contribution of both intersection points of surface with viewing ray and sample points of the volume. We also proposed a framework for time-varying 3D EM datasets. Two assessment tests were conducted by both experts and students. Results showed that our 3D visualization method were effective.

Measurement of Personal Experienced Temperature Variations in Rural Households Using Wearable Monitors: A Pilot Study.

The time-varying data of air temperatures experienced by people in their daily lives is an important basis for studying human thermal sensation, adaptation, comfort, and health. It is also very important for designing targeted strategies to help people reduce uncomfortable experience. In this study, a small (98 mm × 49 mm × 25 mm), lightweight (~100 g), and portable temperature logger with a wide measurement range (−40 to 100 °C) and appropriate accuracy (±0.3 °C precision) was combined with a phone holder that was adapted as an armband sleeve to constitute a wearable monitor. Fourteen monitors were worn by 14 residents in 6 different households in rural Beijing, China, to monitor their personal thermal environment. In the context of having very similar living habits in winter and coping strategies for thermal discomfort, the temperatures that 14 residents experienced exhibited wide ranges and large variations during the two-day test period. The highest and lowest temperatures experienced by residents reached 30.6 and −16.6 °C, respectively. This paper provided new data and evidences about various temperatures experienced by residents, even though they were from the same family and lived together for decades. In terms of methodology, as an exploration, the present study indicated that using personal wearable monitors is a viable method to capture the real experienced thermal environment, which extended the method for collecting data regarding complex experiences in different environments to aid the study of human responses to the real-world thermal environment.

Prediction of the development of islet autoantibodies through integration of environmental, genetic, and metabolic markers.

BackgroundThe Environmental Determinants of the Diabetes in the Young (TEDDY) study has prospectively followed, from birth, children at increased genetic risk of type 1 diabetes. TEDDY has collected heterogenous data longitudinally to gain insights into the environmental and biological mechanisms driving the progression to persistent islet autoantibodies.MethodsWe developed a machine learning model to predict imminent transition to the development of persistent islet autoantibodies based on time‐varying metabolomics data integrated with time‐invariant risk factors (eg, gestational age). The machine learning was initiated with 221 potential features (85 genetic, 5 environmental, 131 metabolomic) and an ensemble‐based feature evaluation was utilized to identify a small set of predictive features that can be interrogated to better understand the pathogenesis leading up to persistent islet autoimmunity.ResultsThe final integrative machine learning model included 42 disparate features, returning a cross‐validated receiver operating characteristic area under the curve (AUC) of 0.74 and an AUC of ~0.65 on an independent validation dataset. The model identified a principal set of 20 time‐invariant markers, including 18 genetic markers (16 single nucleotide polymorphisms [SNPs] and two HLA‐DR genotypes) and two demographic markers (gestational age and exposure to a prebiotic formula). Integration with the metabolome identified 22 supplemental metabolites and lipids, including adipic acid and ceramide d42:0, that predicted development of islet autoantibodies.ConclusionsThe majority (86%) of metabolites that predicted development of islet autoantibodies belonged to three pathways: lipid oxidation, phospholipase A2 signaling, and pentose phosphate, suggesting that these metabolic processes may play a role in triggering islet autoimmunity.

Drivers of CO2-Emissions in Fossil Fuel Abundant Settings: (Pooled) Mean Group and Nonparametric Panel Analyses

The present inquiry addresses the income-environment relationship in oil-producing countries and scrutinizes the further drivers of atmospheric pollution in the respective settings. The existing literature that tests the environmental Kuznets curve hypothesis within the framework of the black-box approaches provides only a bird’s-eye perspective on the long-run income-environment relationship. The aspiration behind this study is making the first step toward the disentanglement of the sources of carbon dioxide emissions, which could be employed in the pollution mitigation policies of this group of countries. Based on the combination of two strands of literature, the environmental Kuznets curve conjecture and the resource curse, the paper at hand proposes an augmented theoretical framework of this inquiry. To approach the research questions empirically, the study employs advanced panel cointegration techniques. To avoid econometric misspecification, the study also employs for the first time a nonparametric time-varying coefficient panel data estimator with fixed effects (NPFE) for the dataset of 37 oil-producing countries in the time interval spanning between 1989 and 2019. The empirical analysis identifies the level of per capita income, the magnitude of oil rents, the share of fossil fuel-based electricity generation in the energy mix, and the share of the manufacturing sector in GDP as essential drivers of carbon dioxide emissions in the oil-rich countries. Tertiarization, on the contrary, leads to a substantial reduction of emissions. Another striking result of this study is that level of political rights and civil liberties are negatively associated with per capita carbon emissions in this group of countries. Furthermore, the study decisively rejects an inverted U-shaped income-emission relationship and validates the monotonically or exponentially increasing impact of average income on carbon dioxide emissions.